JP3640726B2 - Annular member outer periphery inspection apparatus, annular member outer periphery inspection method, and annular member image processing method - Google Patents

Description

但し、ｔ₁ ：ｔ₁ ＝Ｂ_SDC ・Ｌで表わされるしきい値
図２５に示す如く処理対象分割領域の最左端位置ｘ_LEから右方向にサーチし、このサーチ処理で選択された次の位置ｘの垂直射影データｆ（ｘ）と、しきい値Ｔ（ｘ）とを比較し、処理対象となる位置ｘの垂直射影データｆ（ｘ）がしきい値Ｔ（ｘ）より小さいとき、この位置ｘがこの処理対象分割領域に対するピストンリング４の左側面取り位置Ｘ_LCを検出するのに必要な左側面取り基準点Ｘ_LSS でないと判定し、更に右側にある位置ｘを選択し、上述したしきい値Ｔ（ｘ）の計算処理と、垂直射影データｆ（ｘ）としきい値Ｔ（ｘ）との比較処理とを繰り返す。
【００６３】
そして、処理対象として選択された位置ｘの垂直射影データｆ（ｘ）がしきい値Ｔ（ｘ）より大きいか、等しいとき、画像処理装置１１はこの位置ｘがこの処理対象分割領域内におけるピストンリング４の左側面取り位置Ｘ_LCの検出点（左側面取り位置Ｘ_LCをサーチするときの左側面取り基準点）Ｘ_LSS であると判定するとともに、この左側面取り基準点Ｘ_LSS からさらに右側の位置で、垂直射影データｆ（ｘ）の微分関数ｆ’（ｘ）が次式を満たす点（左側最高ピーク位置Ｘ_LMP ）を検出する。
ｆ’（ｘ）≦０
【００６４】
次いで、画像処理装置１１は、上述した左側最高ピーク位置Ｘ_LMP の検出処理と同様な手順で、図２６に示す如く右側初期しきい値Ｔｈ_RPに基づき、この処理対象分割領域にある右側面取り基準点Ｘ_RSS を検出するのに必要な右側初期しきい値Ｔ_{0 RC}を計算する。
【００６５】
その後、画像処理装置１１は処理対象分割領域の最右端位置Ｘ_REから左方向にサーチして、しきい値Ｔ（ｘ）の更新を行いながら、この処理で選択された次の位置ｘの垂直射影データｆ（ｘ）と、しきい値Ｔ（ｘ）とに基づいてピストンリング４の右側面取り位置Ｘ_RCの検出点（右側面取り位置Ｘ_RCをサーチするときの右側面取り基準点）Ｘ_RSS を検出するとともに、この右側面取り基準点Ｘ_RSS からさらに左側の位置ｘで、垂直射影データｆ（ｘ）の微分関数ｆ’（ｘ）がゼロ以下になる点（右側最高ピーク位置Ｘ_RMP ）を検出する。
【００６６】
▲３▼射影データのスムージング、及び▲４▼濃度最小位置検出
次いで、画像処理装置１１は、図２７に示す如く処理対象分割領域の垂直射影データｆ（ｘ）を予め設定された移動平均化幅、例えば２ｍ＋１個の幅で移動平均して、この垂直射影データｆ（ｘ）をスムージング化するとともに、図２８に示す如く分割領域の垂直射影データｆ（ｘ）のうち、左側最高ピーク位置Ｘ_LMP と、右側最高ピーク位置Ｘ_RMP との範囲Ｗ_PT内にある垂直射影データｆ（ｘ）を抽出し、この垂直射影データｆ（ｘ）の値が最小となる位置（最小濃度位置Ｘ_{Min L} ）を検出する。
【００６７】
▲５▼ピストンリング規格寸法よりサーチ範囲決定
その後、画像処理装置１１は次式に示す計算を行って、ピストンリング４の規格に基づき、図２９に示す如く左側面取り基準点Ｘ_LSS と、右側面取り基準点Ｘ_RSS とを基準とした面取り位置サーチ幅Ｗ_PredB を求める。
Ｗ_PredB ＝Ｗ_PB・Ｍａｘ（Ｗ_UR、Ｗ_LR）／Ｗ_Ring
但し、Ｗ_PB：左側面取り基準点Ｘ_LSS と、右側面取り基準点Ｘ_RSS との距離
Ｗ_UR：ピストンリング４の上面側面取り幅
Ｗ_LR：ピストンリング４の下面側面取り幅
Ｗ_Ring：ピストンリング４の幅
【００６８】
▲６▼サーチ範囲の修正
その後、画像処理装置１１は、図３０、図３１に示す如く左側面取り基準点Ｘ_LSS から垂直射影データｆ（ｘ）の最小濃度位置Ｘ_{Min L} までの距離と、面取り位置サーチ幅Ｗ_PredB の値とを比較し、これらのうち、小さい方を左側面取り位置Ｘ_LCのサーチ範囲とし、また右側面取り基準点Ｘ_RSS から垂直射影データｆ（ｘ）の最小濃度位置Ｘ_{Min L} までの距離と、面取り位置サーチ幅Ｗ_PredB の値とを比較し、これらのうち、小さい方を右側面取り位置Ｘ_RCのサーチ範囲とする。
【００６９】
▲７▼サーチ範囲内での射影データの微分
次いで、画像処理装置１１は、左側面取り位置Ｘ_LCのサーチ範囲内にある左側最高ピーク位置Ｘ_LMP から内側に向かって位置ｘを変更しながら、次式に示す計算を行って、これらの各位置ｘに対する微分値ｆ’（ｘ）を求める。
ｆ’（ｘ）＝｛ｆ（ｘ＋Ｄ_IN）−ｆ（ｘ）｝／Ｄ_IN
但し、Ｄ_IN：予め設定されている微分インターバル値
【００７０】
▲８▼微分値ｆ’（ｘ）のサーチ
これらの微分値ｆ’（ｘ）のピークを検出し、それが次式を満たす場合、その位置ｘを左側面取り位置Ｘ_LCとする。
｜ｆ’（ｘ）｜≧Ｔｈ_PS
但し、Ｔｈ_PS：予め設定されている面取り位置検出用のしきい値
また、前記条件を満たす場合、次式に示す計算を行ってしきい値Ｔｈ_PSを変更する。
Ｔｈ_PS＝｜ｆ’（ｘ）｜・Ｔｈ_PSR
但し、Ｔｈ_PSR ：予め設定されている面取り位置検出用ピーク検出化率
【００７１】
その後、画像処理装置１１はこの左側面取り位置Ｘ_LCを求めたときと同様な手順で、右側面取り位置Ｘ_RCを検出する。
【００７２】
以下、画像処理装置１１は残りの各分割領域に対し、垂直射影データｆ（ｘ）の計算処理、左側、右側面取り基準点Ｘ_LSS 、Ｘ_RSS のサーチ処理、左側、右側最高ピーク位置Ｘ_LMP 、Ｘ_RMP の検出処理、垂直射影データｆ（ｘ）のスムージング化処理、最小濃度位置ｘ_{Min L} の検出処理、サーチ範囲の決定処理、サーチ範囲内における左側、右側面取り位置Ｘ_LC、Ｘ_RCの判定処理を行って各分割領域の左側、右側面取り位置Ｘ_LC、Ｘ_RCを求める（ステップＳＴ１１、ＳＴ１２）。
【００７３】
▲９▼面取り位置の修正
１６分割された画像の射影データにより得られた面取り位置について以下のよ処理を行うことにより修正する。
【００７４】
画像処理装置１１は、図３２に示す如く各分割領域の各左側、右側面取り位置Ｘ_LC、Ｘ_RCの左側、右側平均値Ｘ_ALC 、Ｘ_ARC と、左側、右側標準偏差σ_LC、σ_RCとを各々計算した後、各分割領域の左側、右側面取り位置Ｘ_LC、Ｘ_RCの一方、例えば左側面取り位置Ｘ_LCが左側平均値Ｘ_ALC から次式に示す値の範囲内に入っているか否かを計算し、
σ_LC・Ｂ_PRE
但し、Ｂ_PRE ：予め設定されている面取り位置適正評価係数
各左側面取り位置Ｘ_LCのうち、この範囲内に入っていない左側面取り位置Ｘ_LCがあれば、この左側面取り位置Ｘ_LCがバラついていると判定し、図３３に示す如くバラツキが大きいと判定した分割領域の左側面取り位置Ｘ_LCを、その前後にある分割領域の各左側面取り位置Ｘ_LCに基づいて点線に示すように線形補間する。
【００７５】
また、画像処理装置１１は各分割領域の左側面取り位置Ｘ_LCのバラツキ判定処理と同様な手順で、各分割領域の右側面取り位置Ｘ_RCが右側平均値Ｘ_ARC から次式に示す値の範囲内に入っているか否かを計算し、
σ_RC・Ｂ_PRE
但し、Ｂ_PRE ：予め設定されている面取り位置適正評価係数
各右側面取り位置Ｘ_RCのうち、この範囲内に入っていないデータはバラツキが大きいと判断して、バラツキが大きいと判定した分割領域の右側面取り位置Ｘ_RCを、その前後にある分割領域の各右側面取り位置Ｘ_RCに基づいて線形補間する（ステップＳＴ１３）。
【００７６】
＜当たり面検出処理（ステップＳＴ１４からＳＴ２０まで）＞
（１）当たり位置サーチ範囲
画像処理装置１１は、図３４に示す如く、１６分割された各分割領域毎に、次式に示す計算を行って各左側、右側面取り位置Ｘ_LC、Ｘ_RCから内側をサーチして左側、右側当たり位置Ｘ_LE、Ｘ_REを検出するときに必要な非検査領域幅Ｗ_NSB を求めた後、
Ｗ_NSB ＝Ｔｈ_SBR ・（Ｘ_RC−Ｘ_LC）
但し、Ｔｈ_SBR ：当たり位置非検査領域化率
Ｘ_RC：右側面取り位置
Ｘ_LC：左側面取り位置
左側面取り位置Ｘ_LCから右側面取り位置Ｘ_RCまでの範囲のうち、これら左側面取り位置Ｘ_LCから内側の非検査領域幅Ｗ_NSB を除くとともに、右側面取り位置Ｘ_RCから内側の非検査領域幅Ｗ_NSB を除いた範囲を左側、右側当たり位置Ｘ_LE、Ｘ_REを検出するのに必要なサーチ範囲とする。
【００７７】
（２）射影データの区分
その後、画像処理装置１１は、図３５に示す如く、１６分割された各分割領域毎に、サーチ範囲内にある山から山までが小区間となるように、前記サーチ範囲内を複数に分割し、これら各小区間毎に、区間の開始位置、区間の終了位置、区間内の最小値の位置を求めるとともに、全区間で最小値の位置および区間番号、最小値の位置における値（最小値）を求める。
【００７８】
（３）隣接する画面間での小区間同志の連結
その後、図３６に示す如く各分割領域の各小区間を連結して、正しい左側、右側当たり位置Ｘ_LE、Ｘ_REを見つけ出すために、以下の処理を開始する。
【００７９】
まず、画像処理装置１１は、処理対象となる分割領域（処理対象分割領域）の各小区間、例えば図３７に示す如くこの処理対象分割領域が第１、第２、第３小区間に分割されていれば、これら第１小区間中の最小値を与える位置（最小位置）Ｘ_{1 Min} と、第２小区間中の最小値を与える位置（最小位置）Ｘ_{2 Min} と、第３小区間中の最小値を与える位置（最小位置）Ｘ_{3 Min} とを検出する。
【００８０】
そして、画像処理装置１１は処理対象分割領域の下隣りにある分割領域（下隣り分割領域）の各小区間、例えば図３７に示す如くこの下隣り分割領域が第１、第２小区間に分割されていれば、図３８に示す如く処理対象分割領域の第１、第２、第３小区間と、下隣り分割領域の第１、第２小区間とを対応させるとともに、前記下隣り分割領域の垂直射影データｆ（ｘ）の前記処理対象分割領域の各最小位置Ｘ_{1 Min} 、Ｘ_{2 Min} 、Ｘ_{3 Min} における値ｆ（Ｘ_{1 Min} ）、ｆ（Ｘ_{2 Min} ）、ｆ（Ｘ_{3 Min} ）を前記処理対象分割領域の第１、第２、第３小区間に対する下隣り分割領域の第１、第２小区間の評価値とする結合テーブルを作成する。
【００８１】
次いで、画像処理装置１１は処理対象分割領域の上隣りにある分割領域（上隣り分割領域）の各小区間に対し、上述した処理を行って、前記処理対象分割領域の第１、第２、第３小区間に対する上隣り分割領域の各小区間の評価値ｆ（Ｘ_{1 Min} ）、ｆ（Ｘ_{2 Min} ）、ｆ（Ｘ_{3 Min} ）を計算し、結合テーブルを作成する。
【００８２】
以下、画像処理装置１１は残りの分割領域について、同じ計算を行って、各分割領域を処理対象分割領域とする結合テーブルを作成する。
【００８３】
その後、画像処理装置１１は全ての分割領域に対する各評価値ｆ（Ｘ_{1 Min} ）〜ｆ（ｘ_{n Min} ）の各合計値を計算し、これらの各合計値のうち、最小となる合計値の小区間、すなわち正しい左側、右側当たり位置Ｘ_LE、Ｘ_REを含む小区間を選択して、これらの各小区間を連結する（ステップＳＴ１４）。
【００８４】
（４）重み関数の作成
次いで、画像処理装置１１は各分割領域の１つを選択して、これを処理対象分割領域として、左側、右側当たり位置Ｘ_LE、Ｘ_REの検出処理で使用する重み関数の作成に必要なデータ、すなわち図３９に示す如くサーチ範囲内における垂直投影データｆ（ｘ）の最大値と、最小値とを検出するとともに、これら最大値と、最小値とを処理対象分割領域内のライン数で除算して最大位置での濃度値（最大濃度値）ｇ_{P Max L} と、最小位置での濃度値（最小濃度値）ｇ_{P Min L} とを計算するとともに、これら最大濃度値ｇ_{P Max L} と、最小濃度値ｇ_{P Min L} とに基づいて次式に示す計算を行って主濃度レベルＭ_m を求める。
Ｍ_m ＝ｇ_{P Min L} ＋（ｇ_{P Max L} −ｇ_{P Min L} ）・Ｂ_MPR
但し、Ｂ_MPR ：予め設定されている重み関数係数化率
【００８５】
その後、画像処理装置１１はこの主濃度レベルＭ_m と、最大濃度値ｇ_{P Max L} と、最小濃度値ｇ_{P Min L} と、予め設定されている各重み関数係数Ｗ_B （Ｂｏｔｔｏｍ・Ｗｅｉｇｈｔ）、Ｗ_T （Ｔｏｐ・Ｗｅｉｇｈｔ）とに基づいて図４０に示す形状となる重み関数Ｍの重みテーブルを作成する。
【００８６】
（５）サーチ範囲の修正
次いで、画像処理装置１１は検査対象となっているピストンリング４がトップリングであれば、図４１に示す如くサーチ範囲をそのままにし、また検査対象となっているピストンリング４がセカンドリングであれば、図４２に示す如く予め設定されている当たり位置サーチ超過領域Ａ_ESだけ右側面取り位置Ｘ_RCから右側にサーチ範囲を拡げる。
【００８７】
以下、画像処理装置１１は残りの分割領域について、同じ計算を行って、各分割領域毎に、重みテーブルを作成するとともに、サーチ範囲を決定する（ステップＳＴ１５、ＳＴ１６）。
【００８８】
（６）濃度データｇ（ｘ）の微分
その後、画像処理装置１１は各分割領域のうち、処理対象となる分割領域を選択し、この処理対象分割領域内の各ライン毎に、次式に示す計算を行ってピストンリング４の濃度データｇ（ｘ）の微分値ｇ’（ｘ）を求める。
ｇ’（ｘ）＝｛ｇ（ｘ＋Ｄ_IN）−ｇ（ｘ）｝／Ｄ_IN
但し、Ｄ_IN：面取り位置の検出処理で使用した微分インターバル値
【００８９】
（７）当たり位置
次いで、画像処理装置１１は図４３に示す如くこの処理対象分割領域を構成する各小区間のうち、上述した小区間の連結処理で求められた本流となる左側、右側当たり位置Ｘ_LE、Ｘ_REを含む小区間（この図４３では、第２小区間）を選択するとともに、処理対象となっているラインの濃度データｇ（ｘ）のうち、この第２小区間中で最小の濃度を示す位置（最小濃度位置Ｘ_Min ）を検出する。
【００９０】
その後、画像処理装置１１は前記最小濃度位置Ｘ_Min における濃度値（最小濃度値）を使用してサーチ範囲内での最大濃度値ｇ_{Max L} と、最小濃度値ｇ_{Min L} と、最大評価値Ｅ_Max と、最大評価値を与える位置Ｘ_{Max E} を初期化した後、最小濃度位置Ｘ_Min を始点として、位置ｘの値を更新しながら、左側のサーチ範囲をサーチする。
【００９１】
そして、このサーチ処理において、画像処理装置１１は位置ｘに対する濃度データｇ（ｘ）が最大濃度値ｇ_{Max L} より小さければ、この濃度データｇ（ｘ）が最小濃度値ｇ_{Min L} より小さいか否かを判定し、この濃度データｇ（ｘ）が最小濃度値ｇ_{Min L} より小さければ、次式に示す計算を行って最小濃度値ｇ_{Min L} の値を更新する。
ｇ_{Min L} ＝ｇ（ｘ）
【００９２】
また、このサーチ処理において、位置ｘに対する濃度データｇ（ｘ）が最大濃度値ｇ_{Max L} より大きければ、画像処理装置１１は位置ｘにおける微分値ｇ’（ｘ）がピークか否かをチェックし、この微分値ｇ’（ｘ）がピークであれば、この微分値ｇ’（ｘ）と、上述した重みテーブルを参照してこの位置ｘにおける濃度データｇ（ｘ）の値に対応する重み値Ｍに基づいて、次式に示す計算を行い、評価関数Ｅ_V を求め、
Ｅ_V ＝ｇ’（ｘ）・Ｍ
さらに、次式が満たされているか否かを計算し、
Ｅ_V ・Ｃ_EV＞Ｅ_Max
但し、Ｃ_EV：評価関数のダンピング係数
これが満たされていれば、次式に示す計算を行って、最大評価値Ｅ_Max と、最大評価値を与える位置Ｘ_{Max E} を更新する。
Ｅ_Max ＝Ｅ_V
Ｘ_{Max E} ＝ｘ
【００９３】
そして、最小濃度位置Ｘ_Min から左側のサーチ範囲の最左端までサーチされれば、画像処理装置１１はこのとき得られている最大評価値Ｅ_Max と、最大評価値Ｅ_Max を与える位置Ｘ_{Max E} とを左側サーチ範囲の当たり位置（左側当たり位置）Ｘ_LEと判定する。
【００９４】
その後、画像処理装置１１は上述した左側サーチ範囲の当たり位置Ｘ_LEを求めた手順と同じ手順で、前記最小濃度位置Ｘ_Min における濃度値（最小濃度）を使用してサーチ範囲内での最大濃度値ｇ_{Max L} と、最小濃度値ｇ_{Min L} と、最大評価値Ｅ_Max と、最大評価値を与える位置Ｘ_{Max E} を初期化した後、位置ｘの値を更新しながら、最小濃度位置Ｘ_Min から右側のサーチ範囲をサーチして右側サーチ範囲の当たり位置（右側当たり位置）Ｘ_REを計算する。
【００９５】
以下、画像処理装置１１は、この処理対象分割領域内にある残りの各ラインについて、左側当たり位置Ｘ_LEと、右側当たり位置Ｘ_REとを計算する。
【００９６】
そして、この処理対象分割領域内にある各ライン毎の左側当たり位置Ｘ_LEと、右側当たり位置Ｘ_REとの計算が終了したとき、画像処理装置１１は残っている分割領域を順次、選択してこれを処理対象分割領域とした後、上述した処理を行って各ライン毎に、左側当たり位置Ｘ_LEと、右側当たり位置Ｘ_REとを計算する（ステップＳＴ１７）。
【００９７】
（８）当たり位置の修正（固定レベルサーチ）
次いで、画像処理装置１１は、よごれ等による飛びを修正するために、図４４に示す如く全てのラインのうち、処理対象となるラインを選択し、このライン（注目ライン）に対する左側当たり位置Ｘ_LEに対する濃度データｇ（ｘ）と、この注目ラインの前後、２ラインに対する左側当たり位置Ｘ_LEに対する濃度データｇ（ｘ）とに基づき、最大の濃度データｇ（ｘ）と、最小の濃度データｇ（ｘ）とを除く３つの濃度データｇ（ｘ）の平均値（左側平均濃度データ）Ｔ_LAを計算する。
【００９８】
その後、画像処理装置１１は、この左側平均濃度データＴ_LAが次式を満たすか否かをチェックし、
Ｔ_LA＜Ｂ_{S Min L} ＋ｇ_{L Min L}
但し、Ｂ_{S Min L} ：予め設定されている２値化サーチ最小値
ｇ_{L Min L} ：注目ラインを含む小区間の最小濃度値
この条件が満たされていれば、次式に基づいて左側平均濃度データＴ_LAを修正してこれが小さくなり過ぎないようにする。
Ｔ_LA＝Ｂ_{S Min L} ＋ｇ_{L Min L}
【００９９】
その後、画像処理装置１１は注目ライン上で、左側平均濃度データＴ_LAより大きな値を持つ濃度データｇ（ｘ）の位置ｘを検出して、この位置ｘを正しい左側当たり位置Ｘ_LEであると判定し、注目ライン上の左側当たり位置Ｘ_LEを修正する。
【０１００】
以下、画像処理装置１１は残りの各ラインのうち、次のラインを順次、選択してこれを注目ラインにしながら、上述した注目ライン上の左側当たり位置Ｘ_LEの修正を行う。
【０１０１】
次いで、画像処理装置１１は全てのライン上にある右側当たり位置Ｘ_REについて、上述した左側当たり位置Ｘ_LEの修正処理と同じ処理を行って、各ラインの右側当たり位置Ｘ_REを修正する（ステップＳＴ１８）。
【０１０２】
（９）当たり位置の修正（ジャンプ修正）
その後、画像処理装置１１は全分割領域内にある各ラインの１つを選択して、これを注目ラインとした後、この注目ラインの左側当たり位置Ｘ_LEと、この注目ラインの前後、１ラインの左側当たり位置Ｘ_LEとの差を計算するとともに、これらの差が正のとき、すなわち図４５に示す如く前後ラインの左側当たり位置Ｘ_LEより注目ラインの左側当たり位置Ｘ_LEが外側にあるとき、次式が満たされているか否かをチェックし、
ａ_1b＞Ｌｉｍｉｔ_JP
ａ_2a＞Ｌｉｍｉｔ_JP
但し、Ｌｉｍｉｔ_JP：予め設定されている当たり位置ジャンプリミット値 （外側）
ａ_1b：注目ラインより前にあるラインの当たり位置Ｘ_LEと、注目ラインの当たり位置Ｘ_LEとの差
ａ_2a：注目ラインより後にあるラインの当たり位置Ｘ_LEと、注目ラインの当たり位置Ｘ_LEとの差
これらが同時に満たされているとき、前後ラインの左側当たり位置Ｘ_LEより注目ラインの左側当たり位置Ｘ_LEがジャンプした位置にあると判定して、前後ライン上にある各左側当たり位置Ｘ_LEに基づいて注目ライン上にある左側当たり位置Ｘ_LEを直線補間して、これを修正する。
【０１０３】
また、図４６に示す如く前後ラインの左側当たり位置Ｘ_LEより注目ラインの左側当たり位置Ｘ_LEが内側にあるとき、画像処理装置１１は次式が満たされているか否かをチェックし、
ｂ_1b＞Ｌｉｍｉｔ_JM
ｂ_2a＞Ｌｉｍｉｔ_JM
但し、Ｌｉｍｉｔ_JM：予め設定されている当たり位置ジャンプリミット値 （内側）
ｂ_1b：注目ラインより前にあるラインの当たり位置Ｘ_LEと、注目ラインの当たり位置Ｘ_LEとの差
ｂ_2a：注目ラインより後にあるラインの当たり位置Ｘ_LEと、注目ラインの当たり位置Ｘ_LEとの差
これらが同時に満たされているとき、前後ラインの左側当たり位置Ｘ_LEに基づいて注目ラインの左側当たり位置Ｘ_LEを直線補間してこれを修正する。
【０１０４】
以下、画像処理装置１１は残りの各ラインのうち、次のラインを順次、選択してこれを注目ラインにしながら、上述した注目ライン上の左側当たり位置Ｘ_LEの修正を行い、これら各左側当たり位置Ｘ_LEを結ぶ線を滑らかにする。
【０１０５】
次いで、画像処理装置１１は上述した左側当たり位置Ｘ_LEの修正処理に使用した手順と同様な手順で全分割領域内にある各ラインの右側当たり位置Ｘ_REについて、直線補間方式でこれを修正し、右側当たり位置Ｘ_REを結ぶ線を滑らかにする（ステップＳＴ１９）。
【０１０６】
＜当たり位置スムージング＞
次いで、当たり位置のデータから大きくジャンプしている箇所を探し、スムージングを行う。画像処理装置１１は全分割領域を構成する各ラインの１つを選択してこれを注目ラインとした後、図４７に示す如く注目ラインの左側当たり位置Ｘ_LEと、この注目ラインから一定ライン数（ジャンプ位置検出インターバルＭ_{1 IN}）だけ下方に離れた位置にあるライン（参照ライン）の左側当たり位置Ｘ_LEとの差Δｂを計算し、この差Δｂがそれまでの差Δｂより大きいか否かを判定する。
【０１０７】
そして、注目ラインの差Δｂがそれまでの差Δｂより大きいとき、画像処理装置１１は図４８に示す如く注目ラインから一定ライン数（ジャンプ位置検出インターバルＭ_{2 IN}）だけ下方に離れた位置にあるライン（再参照ライン）の左側当たり位置Ｘ_LEと、前記参照ラインから一定ライン数（ジャンプ位置検出インターバルＭ_{2 IN}）だけ上方に離れた位置にあるライン（再参照ライン）の左側当たり位置Ｘ_LEとの差Δｗを計算する。
【０１０８】
そして、この差Δｗがそれまでの処理によって得られている差Δｗの最大値より大きいとき、画像処理装置１１は前記差Δｂをこれまでの最大値（最大差）Ｍａｘ（Δｂ）として記憶する。
【０１０９】
以下、画像処理装置１１は、注目ラインを１ラインずつずらしながら、残りの各ラインについて、上述した差Δｂ、Δｗの計算処理を行い、新たに得られた差Δｂがそれまでの差Δｂの最大値より大きいとき、新たに得られた差Δｗがそれまでの差Δｗより大きいか否かを判定し、新たに得られた差Δｗがそれまでの差Δｗより大きいとき、新たに得られた差Δｂによって最大差Ｍａｘ（Δｂ）を更新する。
【０１１０】
そして、全てのラインに対する最大差Ｍａｘ（Δｂ）が得られたとき、画像処理装置１１はこの最大差Ｍａｘ（Δｂ）が次式を満たすか否かを計算する。
Ｍａｘ（Δｂ）≧Ｌ_JS
但し、Ｌ_JS：予め設定されている最低ジャンプ長
【０１１１】
そして、この条件が満たされていれば、画像処理装置１１は図４９に示す如く各ラインのうち、前記最大差Ｍａｘ（Δｂ）となったラインを含む左側当たり位置Ｘ_LEをジャンプ点と判定した後、次式に示す計算を行って、このジャンプ点を含む各左側当たり位置Ｘ_LEを結んだ線のジャンプ部分をスムージングするのに必要なスムージング長さＬ_Smを求め、前記ジャンプ点から前記スムージング長さＬ_Smだけ上のラインをスムージング処理の開始点Ｙ_S と決定するとともに、前記ジャンプ点に対応する参照点から前記スムージング長さＬ_Smだけ下のラインをスムージング処理の終了点Ｙ_Enと決定する。
Ｌ_Sm＝Ｍａｘ（Δｂ）
・｛Ｃ_{Sm F}・Ｄ／（Ｂ−ｂ_Up−ｂ_Lo）｝
但し、Ｃ_{Sm F}：予め設定されているスムーズ長算出係数
Ｄ：検査対象となるピストンリング４の直径（ｍｍ）
Ｂ：検査対象となるピストンリング４の厚さ（ｍｍ）
ｂ_Up：検査対象となるピストンリング４の上面取り（ｍｍ）
ｂ_Lo：検査対象となるピストンリング４の下面取り（ｍｍ）
【０１１２】
その後、画像処理装置１１は、図５０に示す如く開始点Ｙ_S から予め設定されているスムージング長さＭ_{3 Sm}だけ上のラインまでの区間を上側境界処理区間と決定し、この上側境界処理区間内にある各ラインの左側当たり位置Ｘ_LEを結んだ線を関数ｆ（ｙ）とするとともに、終了点Ｙ_Enから予め設定されているスムージング長さＭ_{3 Sm}だけ下のラインまでの区間を下側境界処理区間と決定し、この下側境界処理区間内にある各ラインの左側当たり位置Ｘ_LEを結んだ線を関数ｆ（ｙ）とし、これら上側境界処理区間内にある各ラインの左側当たり位置Ｘ_LEを示す関数ｆ（ｙ）と、下側境界処理区間内にある各ラインの当たり位置Ｘ_LEを示す関数ｆ（ｙ）とに対して最小２乗法による直線近似を行って近似直線にする。
【０１１３】
次いで、画像処理装置１１は、図５１に示す如くスムーズ関数を使用して上側境界処理区間の近似直線の下端と、下側境界処理区間の近似直線の上端とを滑らかに結ぶ３次曲線を作成して各ラインの左側当たり位置Ｘ_LEを結ぶ線分のジャンプ部分を滑らかにし、これら左側の当たり位置Ｘ_LEを結ぶ線全体を連続化させる。
【０１１４】
そして、左側当たり位置Ｘ_LEのスムージング処理が終了すれば、画像処理装置１１は、上述した左側当たり位置Ｘ_LEのスムージング処理と同様な手順で、各ラインの右側当たり位置Ｘ_REに対するスムージング処理を行って、右側当たり位置Ｘ_REを結んだ線のジャンプ部分を滑らかにして、これら右側当たり位置Ｘ_REを結んだ線全体を連続化させる（ステップＳＴ２０）。
【０１１５】
＜判定処理（ステップＳＴ２１からＳＴ２３まで）＞
（１）当たり幅の検査
画像処理装置１１は、各ライン毎に左側当たり位置Ｘ_LEを結んだ線と右側当たり位置Ｘ_REを結んだ線との距離（当たり幅）Ｗ₁ を計算するとともに、図５２に示す如く、これらの当たり幅Ｗ₁ のうち、ピストンリング４の合口部（フリーギャップ部）から所定長さＬ_T の部分の当たり幅Ｗ₁ が次式に示す条件を満たすか否かを計算する。
Ｗ₁ ≧Ｌｉｍｉｔ_LoBT
Ｗ₁ ≦Ｌｉｍｉｔ_UpBT
但し、Ｌｉｍｉｔ_LoBT：合口部の当たり幅下限値
Ｌｉｍｉｔ_UpBT：合口部の当たり幅上限値
【０１１６】
その後、画像処理装置１１は、合口部から所定長さＬ_T の部分を除いた部分の当たり幅Ｗ₁ が次式に示す条件を満たすか否かを計算する。
Ｗ₁ ≧Ｌｉｍｉｔ_LoB
Ｗ₁ ≦Ｌｉｍｉｔ_UpB
但し、Ｌｉｍｉｔ_LoB ：当たり幅の下限値
Ｌｉｍｉｔ_UpB ：当たり幅の上限値
【０１１７】
そして、これらの各条件が同時に満たされていれば、画像処理装置１１はこのピストンリング４の当たり幅Ｗ₁ が良好であると判定し、またこれらの各条件の少なくともいずれか１つが満たされていなければ、このピストンリング４の当たり幅Ｗ₁ が不良になっていると判定する（図９のステップＳＴ２１）。
【０１１８】
（２）残り幅の検査
次いで、このピストンリング４がトップリングであれば、画像処理装置１１は、図５３に示す如く各ライン毎に、左側面取り位置（上側面取り位置）Ｘ_LCと当たり面の左側当たり位置Ｘ_LEとの距離（上側残り幅）Ｗ_{2 UpRe}と、右側面取り位置（下側面取り位置）Ｘ_RCと当たり面の右側当たり位置Ｘ_RCとの距離（下側残り幅）Ｗ_{2 LoRe}とを計算するとともに、これら上側残り幅Ｗ_{2 UpRe}と下側残り幅Ｗ_{2 LoRe}とが次式を満たしているか否かを計算する。
Ｗ_{2 UpRe}≦Ｌｉｍｉｔ_WURT
Ｗ_{2 LoRe}≧Ｌｉｍｉｔ_WLRT
但し、Ｌｉｍｉｔ_WURT：トップリングに対する上側残り幅上限値
Ｌｉｍｉｔ_WLRT：トップリングに対する下側残り幅下限値
【０１１９】
そして、これらの各条件が同時に満たされていれば、画像処理装置１１は、このピストンリングの上残り幅Ｗ_{2 UpRe}と下側残り幅Ｗ_{2 LoRe}とが良好であると判定し、またこれらの各条件の少なくともいずれか一方が満たされていなければ、上残り幅Ｗ_{2 UpRe}を下側残り幅Ｗ_{2 LoRe}にするとともに、下側残り幅Ｗ_{2 LoRe}を上残り幅Ｗ_{2 UpRe}にして、上述した上側残り幅Ｗ_{2 UpRe}の良否判定処理、下側残り幅Ｗ_{2 LoRe}の良否判定処理を行い、これらの良否判定処理によって各条件が満たされている場合には、ピストンリング４の上下を逆にすれば、残り幅について良品となる半良品と判定する。
【０１２０】
また、このような処理を行っても、上述した各条件が同時に満たされていない場合には、画像処理装置１１はこのピストンリング４の上残り幅Ｗ_{2 UpRe}、または下側残り幅Ｗ_{2 LoRe}の少なくとも一方が不良になっていると判定する。
【０１２１】
また、このピストンリング４がセカンドリングであれば、画像処理装置１１は図５４に示す如く各ライン毎に、下側面取り位置Ｘ_RCと当たり面の右側当たり位置Ｘ_REの距離（下側残り幅）Ｗ_{2 LoRe}を計算するとともに、この下側残り幅Ｗ_{2 LoRe}が次式を満たしているか否かを計算する。
Ｗ_{2 LoRe}≦Ｌｉｍｉｔ_WURS
但し、Ｌｉｍｉｔ_WURS：セカンドリングに対する下側残り幅上限値
【０１２２】
そして、この条件が満たされていれば、画像処理装置１１はこのピストンリング４の下側残り幅Ｗ_{2 LoRe}が良好であると判定し、またこの条件が満たされていなければ、このピストンリング４の下側残り幅Ｗ_{2 LoRe}が不良になっていると判定する（ステップＳＴ２２）。
【０１２３】
（３）当たり切れ
次いで、画像処理装置１１は各ライン毎に、上述した左側、右側当たり位置Ｘ_LE、Ｘ_REの検出処理で得られた左側当たり位置Ｘ_LEの評価関数Ｅ_V と、右側当たり位置Ｘ_REの評価関数Ｅ_V との平均値Ｅ_AVを計算するとともに、次式に基づいて当たり切れ（当たりが切れている部分をいう。当たり切れ部分は周囲に比べて明るくなる。）の判定に必要なしきい値Ｔｈ₂ を求める。
Ｔｈ₂ ＝Ｅ_AV・Ｅ_{V LoRa}
但し、Ｅ_{V LoRa}：経験によって予め設定されているコントラスト下限比率
【０１２４】
その後、画像処理装置１１は各ライン毎に、上述した左側、右側当たり位置Ｘ_LE、Ｘ_REの検出処理で得られた最小濃度値ｇ_{Min L} のうち、最も小さい値の最小濃度値ｇ_{Min L} を選択し、これを極小濃度値ｇ_{Min ML}として次式に示す計算を行い、当たり切れの判定に必要なしきい値Ｔｈ₃ を求めた後、
Ｔｈ₃ ＝ｇ_{Min ML}・Ｌ_{UMin R}
但し、Ｌ_{UMin R}：予め設定されている極小値上限比率
各ライン毎に、最小濃度値ｇ_{Min L} が次式を満たしているか否かを計算する。
ｇ_{Min L} ＞Ｔｈ₃
【０１２５】
その後、画像処理装置１１は各ライン毎に、左側当たり位置Ｘ_LEの評価関数Ｅ_V が次式に示す条件が満たされているか否かを計算するとともに、
Ｅ_V ＜Ｌｉｍｉｔ_EVL 、またはＥ_V ＜Ｔｈ₂
但し、Ｌｉｍｉｔ_EVL ：予め設定されている評価関数の下限値
右側当たり位置Ｘ_REの評価関数Ｅ_V が次式に示す条件が満たされているか否かを計算する。
Ｅ_V ＜Ｌｉｍｉｔ_EVL 、またはＥ_V ＜Ｔｈ₂
【０１２６】
そして、これらの各条件の全てに当てはまらないとき、画像処理装置１１はこのピストンリング４が当たり切れを起こしていないと判定し、また前記各条件の１つでも当てはまるとき、このピストンリング４に当たり切れがあると判定する。
【０１２７】
（４）当たり中心位置
次いで、画像処理装置１１は図５５に示す如く各ライン毎に、左側当たり位置Ｘ_LEと、右側当たり位置Ｘ_REとの平均値を計算して、当たり面の中心位置を求めるとともに、図５６に示す如く上側面取り位置Ｘ_LEと、当たり面の中心位置との距離Ｌ_C を計算し、この距離Ｌ_C が次式を満たしているか否かを計算する。
Ｌｉｍｉｔ_{Lo C}＜Ｌ_C ＜Ｌｉｍｉｔ_{Up C}
但し、Ｌｉｍｉｔ_{Lo C}：予め設定されている当たり位置中心の下限値
Ｌｉｍｉｔ_{Up C}：予め設定されている当たり位置中心の上限値
【０１２８】
そして、この条件が満たされていれば、画像処理装置１１はこのピストンリング４の当たり位置の中心が良好であると判定し、またこの条件が満たされていなければ、このピストンリング４の当たり位置の中心が不良になっていると判定する。
【０１２９】
その後、画像処理装置１１は図５７に示す如く各ラインの距離Ｌ_C を比較して最大の値を持つ距離Ｌ_C を選択し、これを最大値Ｌ_{C Max} とするとともに、前記各ラインの距離Ｌ_C を比較して最小の値を持つ距離ＬC を選択し、これを最小値Ｌ_{C Min} とした後、次式に示す計算を行って当たり位置中心の振れ幅Ａ_L を求めた後、
Ａ_L ＝Ｌ_{C Max} −Ｌ_{C Min}
次式に示す条件が満たされているか否かを計算する。
ＡL ＜Ｌｉｍｉｔ_{Up CFl}
但し、Ｌｉｍｉｔ_{Up CFl}：予め設定された当たり位置中心の振れ幅上限値
【０１３０】
そして、この条件が満たされていれば、画像処理装置１１はこのピストンリング４の当たり位置の中心の振れ幅Ａ_L が良好であると判定し、またこの条件が満たされていなければ、このピストンリング４の当たり位置の中心の振れ幅Ａ_L が不良になっていると判定する。
【０１３１】
その後、画像処理装置１１は、上述した全ての検査で良好と判定されたピストンリング４に対して良品を指示する制御信号を生成し、また上述した各検査のいずれかで不良と判定されたピストンリング４に対して不良品を指示する制御信号を生成し、これを撮像機構２に供給してピストンリング４の選別を行わせる（ステップＳＴ２３）。
【０１３２】
＜表示画像作成処理（ステップＳＴ２４）＞
そして、パーソナルコンピュータ装置１３によって検査結果表示画面が指定されているとき、画像処理装置１１は、上述した検査結果によって得られた各データを使用して図５８に示す如く現在検査中のピストンリング４の画像を処理して見易くしたピストン画像と、このピストンリング４を検査して良品、不良品、半良品のいずれかに分類した良否判定結果と、このピストンリング４にある当たり面の幅不良の有無およびその位置と、このピストンリング４にある当たり面の当たり切れの有無およびその位置と、このピストンリング４にある当たり面と面取り位置との間の距離（残り幅）の不良有無およびその位置と、このピストンリング４にある当たり面の左側に対するスムージング処理の有無およびその位置と、このピストンリング４にある当たり面の右側に対するスムージング処理の有無およびその位置と、それまでに検査したピストンリング４の総数と、それまでに検査したピストンリング４中の良品総数と、それまでに検査したピストンリング４中の半良品総数と、それまでに検査したピストンリング４中の不良品総数とを表示する画面データを生成し、これを各モニタ装置８、１２上に表示させる。
【０１３３】
また、パーソナルコンピュータ装置１３によって、検査状況表示画面が指定されているとき、画像処理装置１１は上述した検査結果によって得られた各データを使用して図５９に示す如く現在検査中のピストンリング４の画像を処理して見易くしたピストン画像と、このピストンリング４を検査して良品、不良品、半良品のいずれかに分類した良否判定結果と、このピストンリング４にある当たり面の幅不良の有無およびその位置と、このピストンリング４にある当たり面の当たり切れの有無およびその位置と、このピストンリング４にある当たり面と面取り位置との間の距離（残り幅）の不良有無およびその位置と、このピストンリング４にある当たり面の左側に対するスムージング処理の有無およびその位置と、このピストンリング４にある当たり面の右側に対するスムージング処理の有無およびその位置と、このピストンリング４の画像の輝度と、このピストンリング４の画像のコントラストと、このピストンリング４の画像のエッジ強度と、このピストンリング４の画像の鮮鋭度とを表示する画面データを生成し、これを各モニタ装置８、１２上に表示させる（ステップＳＴ２４）。
【０１３４】
このようにこの実施の形態においては、撮像機構によってピストンリング４を取り込み、これを回転させながら、外周部分を撮像するとともに、制御処理機構３によって前記撮像動作で得られた画像を処理して、ピストンリング４のフリーギャップ検出、外側エッジ検出、面取り位置検出、当たり面検出、良否判定、表示画像作成などを行い、前記ピストンリング４の良否を判定するようにしたので、ＩＴＶカメラとして使用されるラインセンサ１１によってピストンリング外周の画像を取り込んだとき、ピストンリング画像がゆがんでも、またピストンリング画像のコントラストが小さくても、高い精度でピストンリング４の良否を判定することができる。
【０１３５】
【発明の効果】
以上説明したように、請求項１乃至請求項３記載の発明では、カメラから環状部材（ピストンリング）外周の画像を取り込んだとき、環状部材（ピストンリング）画像が歪んでいる場合にあっても、また環状部材（ピストンリング）画像のコントラストが小さくても、高い精度で環状部材（ピストンリング）の良否を判定することができる。
【０１３６】
また、請求項４記載の発明では、検査対象となる環状部材（ピストンリング）の画像処理を容易にすることができるとともに、環状部材（ピストンリング）の画像を見易くすることができる。
【０１３７】
また、請求項５記載の発明では、検査対象となる環状部材（ピストンリング）の画像処理の処理精度を高めることができる。
【０１３８】
さらに、請求項６記載の発明では、検査対象となる環状部材（ピストンリング）の画像中の検出したい部分を高い精度で検出することができる。
【図面の簡単な説明】
【図１】本発明による環状部材の外周検査装置および画像処理方法の実施の形態を示すブロック図である。
【図２】図１に示す環状部材の外周検査装置によって検査されるピストンリングの一例を示す側面図である。
【図３】図１に示す環状部材の外周検査装置の検査項目の一例を示す模式図である。
【図４】図１に示す検査ステーション装置のピストンリング画像撮像動作例を示す側面図である。
【図５】図１に示す検査ステーション装置のピストンリング画像撮像動作例を示す上面図である。
【図６】図１に示す撮像機構から出力される画像信号の内容例を示す模式図である。
【図７】図１に示す環状部材の外周検査装置の検査動作例を示すフローチャートである。
【図８】図１に示す環状部材の外周検査装置の検査動作例を示すフローチャートである。
【図９】図１に示す環状部材の外周検査装置の検査動作例を示すフローチャートである。
【図１０】図１に示す制御処理機構のフリーギャップ検出処理で行われる有効画像切出し動作例を示す模式図である。
【図１１】図１に示す制御処理機構のフリーギャップ検出処理で行われる水平射影処理領域の切出し動作例を示す模式図である。
【図１２】図１に示す制御処理機構のフリーギャップ検出処理で行われる水平射影計算動作例を示す模式図である。
【図１３】図１に示す制御処理機構のフリーギャップ検出処理で行われる外側エッジサーチのしきい値計算例を示す模式図である。
【図１４】図１に示す制御処理機構のフリーギャップ検出処理で行われる外側エッジサーチ動作例を示す模式図である。
【図１５】図１に示す制御処理機構のフリーギャップ検出処理によって検出されたフリーギャップ例を示す模式図である。
【図１６】図１に示す制御処理機構のフリーギャップ検出処理によって検出されたフリーギャップ例を示す模式図である。
【図１７】図１に示す制御処理機構の外側エッジ検出処理で検出される外側エッジ例を示す模式図である。
【図１８】図１に示す制御処理機構の外側エッジ検出処理で行われるしきい値の計算動作例を示す模式図である。
【図１９】図１に示す制御処理機構の外側エッジ検出処理で行われる左側の外側エッジサーチ動作例を示す模式図である。
【図２０】図１に示す制御処理機構の外側エッジ検出処理で行われる左側の外側エッジの決定例を示す模式図である。
【図２１】図１に示す制御処理機構の外側エッジ検出処理で行われる右側の外側エッジサーチ動作例を示す模式図である。
【図２２】図１に示す制御処理機構の外側エッジ検出処理で決定された処理領域の一例を示す模式図である。
【図２３】図１に示す制御処理機構の面取り位置検出処理で行われる処理対象領域の分割例を示す模式図である。
【図２４】図１に示す制御処理機構の面取り位置検出処理で行われる垂直射影データの計算動作例を示す模式図である。
【図２５】図１に示す制御処理機構の面取り位置検出処理で行われる面取り位置と、ピーク位置とのサーチ動作例を示す模式図である。
【図２６】図１に示す制御処理機構の面取り位置検出処理で得られた面取り位置と、ピークとの一例を示す模式図である。
【図２７】図１に示す制御処理機構の面取り位置検出処理で行われる移動平均処理の一例を示す模式図である。
【図２８】図１に示す制御処理機構の面取り位置検出処理で行われる最小位置の検出動作例を示す模式図である。
【図２９】図１に示す制御処理機構の面取り位置検出処理で得られた面取り位置の仮のサーチ範囲例を示す模式図である。
【図３０】図１に示す制御処理機構の面取り位置検出処理で行われる面取り位置のサーチ範囲の決定例を示す模式図である。
【図３１】図１に示す制御処理機構の面取り位置検出処理で行われる面取り位置のサーチ範囲の決定例を示す模式図である。
【図３２】図１に示す制御処理機構の面取り位置検出処理で得られた面取り位置の修正動作例を示す模式図である。
【図３３】図１に示す制御処理機構の面取り位置検出処理で行われる面取り位置の一例を示す模式図である。
【図３４】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のサーチ範囲の一例を示す模式図である。
【図３５】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のサーチ範囲の分割例を示す模式図である。
【図３６】図１に示す制御処理機構の当たり面検出処理で行われる小区間同士の連結例を示す模式図である。
【図３７】図１に示す制御処理機構の当たり面検出処理で行われる小区間同士の連結動作で使用される評価値の計算動作例を示す模式図である。
【図３８】図１に示す制御処理機構の当たり面検出処理で行われる小区間同士の連結動作で使用される総合テーブルの一例を示す図である。
【図３９】図１に示す制御処理機構の当たり面検出処理で行われる重み関数の作成動作例を示す模式図である。
【図４０】図１に示す制御処理機構の当たり面検出処理で作成される重みテーブルの一例を示す模式図である。
【図４１】図１に示す制御処理機構の当たり面検出処理で行われるトップリングに対する当たり位置のサーチ範囲例を示す模式図である。
【図４２】図１に示す制御処理機構の当たり面検出処理で行われるセカンドリングに対する当たり位置のサーチ範囲例を示す模式図である。
【図４３】図１に示す制御処理機構の当たり面検出処理で行われるサーチ範囲の分割動作例を示す模式図である。
【図４４】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置の修正動作例を示す模式図である。
【図４５】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置の修正動作例を示す模式図である。
【図４６】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置の修正動作例を示す模式図である。
【図４７】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のスムージング箇所の判定動作例を示す模式図である。
【図４８】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のスムージング箇所の判定動作例を示す模式図である。
【図４９】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のスムージング対象区間の一例を示す模式図である。
【図５０】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のスムージング動作の境界部分の一例を示す模式図である。
【図５１】図１に示す制御処理機構の当たり面検出処理で行われる当たり位置のスムージング動作例を示す模式図である。
【図５２】図１に示す制御処理機構の判定処理で処理されるピストンリングの一例を示す斜視図である。
【図５３】図１に示す制御処理機構の判定処理で行われるトップリングに対する残り幅の判定動作例を示す模式図である。
【図５４】図１に示す制御処理機構の判定処理で行われるセカンドリングに対する残り幅の判定動作例を示す模式図である。
【図５５】図１に示す制御処理機構の判定処理で行われる当たり位置の中心検出動作例を示す模式図である。
【図５６】図１に示す制御処理機構の判定処理で行われる当たり位置中心の判定動作例を示す模式図である。
【図５７】図１に示す制御処理機構の判定処理で行われる当たり位置の振れ幅の判定動作例を示す模式図である。
【図５８】図１に示す制御処理機構の表示画像作成処理で表示される表示画面の一例を示す模式図である。
【図５９】図１に示す制御処理機構の表示画像作成処理で表示される表示画面の一例を示す模式図である。
【符号の説明】
１ 環状部材の外周検査装置
２ 撮像機構
３ 制御処理機構
４ ピストンリング（環状部材）
５ 搬送制御装置
６ 検査ステーション装置
７ 搬送装置
８ 画像モニタ装置
９ 照明電源装置
１０ ライトガイド
１５ 保持機構
１１ ラインセンサ（カメラ）
１１ 画像処理装置
１２ 画像モニタ装置
１３ パーソナルコンピュータ装置
１４ シリアルプリンタ装置[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an annular member outer periphery inspection device, an annular member outer periphery inspection method, and an annular member image processing method for inspecting the outer periphery of a non-circular annular member such as a piston ring used in an internal combustion engine.
[0002]
[Prior art]
Conventionally, there is an outer periphery inspection process as one of inspection processes of piston rings used in internal combustion engines. In this outer periphery inspection process, the piston ring is actually attached to the piston, the piston ring is slid in the cylinder, and the state of wear generated on the outer periphery of the piston ring is confirmed by visual inspection or the like. A non-defective product / defective product is sorted by determining a product that is uniform over the entire circumference of the ring and has an appropriate width as a good product.
[0003]
[Problems to be solved by the invention]
However, since the conventional visual inspection described above is inefficient and high inspection accuracy cannot be desired, an image of the outer periphery of the piston ring is captured using an ITV camera or the like, and the image information obtained thereby is included. Based on this, it is strongly desired to develop an inspection method for determining the quality.
[0004]
However, in such an inspection method based on image processing, when an image of the piston ring outer periphery is captured by the ITV camera, the piston ring image is distorted and it is difficult to determine the wear width (contact surface) of the piston ring. At present, there is no high-precision inspection device.
[0005]
In view of the above circumstances, in the present invention, when an image of the outer periphery of an annular member such as a piston ring is captured by an ITV camera or the like, even if the annular member (piston ring) image is distorted, the contrast of the annular member (piston ring) image is small. However, it is an object of the present invention to provide an annular member outer periphery inspection device, an annular member outer periphery inspection method, and an annular member image processing method that can determine the quality of an annular member (piston ring) with high accuracy.
[0006]
[Means for Solving the Problems]
In order to achieve the above object, the present invention provides an annular member outer periphery inspection apparatus according to claim 1, wherein the annular member side to be inspected is rotated or separated from the annular member by a predetermined distance. The outer peripheral portion of the annular member by revolving and imaging the outer peripheral portion of the annular member, and generating an image signal of the outer peripheral portion of the annular member, and the outer periphery of the annular member based on the image signal of the annular member obtained by the imaging mechanism A control processing mechanism for inspecting, and the control processing mechanism includes processing target region determination means for determining a processing target region based on an image signal of the annular member, and each line in the determined processing target region. Image specifying means for detecting the outer edge of the annular member for each time and specifying the annular member image; chamfer position detecting means for detecting the chamfering position of the annular member by searching the specified annular member image; A contact surface detecting means for detecting a contact surface by searching a range corresponding to the chamfered position, and a determination means for determining the quality of the annular member based on the state of the detected contact surface. It is a feature.
[0007]
According to the above-described configuration, the annular member side to be inspected is rotated by the imaging mechanism or the camera side arranged at a predetermined distance from the annular member is driven to revolve, and the outer peripheral portion of the annular member is imaged. After the image signal of the outer peripheral part of the annular member is generated and the processing target area is determined based on the image signal of the annular member obtained by the imaging mechanism by the control processing mechanism, The outer edge of the annular member is detected for each line, the annular member image is specified, and the inside of the annular member image is searched to detect the chamfering position of the annular member, and then corresponds to the chamfering position of the annular member image. The contact area is searched and a contact surface is detected, and whether the annular member is good or bad is determined based on the state of the contact surface.
[0008]
The annular member outer periphery inspection device according to claim 2 is the annular member outer periphery inspection device according to claim 1, wherein the annular member is a piston ring, and the processing target region determining means detects a free gap of the piston ring to determine a processing target region. It is characterized by deciding.
[0009]
According to the above configuration, after the piston ring free gap is detected and the processing target region is determined based on the image signal of the piston ring, the outer edge of the piston ring is detected for each line in the processing target region. The piston ring image is specified, and the piston ring image is further searched to detect the chamfering position of the piston ring, and then the range corresponding to the chamfering position of the piston ring image is searched to detect the contact surface. Then, based on the state of the contact surface, whether the piston ring is good or bad is determined.
[0010]
Further, in the annular member outer periphery inspection method according to claim 3, the outer peripheral portion of the annular member is rotated by rotating the annular member side to be inspected or revolving the camera side disposed at a predetermined distance from the annular member. To generate an image signal, determine a processing target region based on the image signal of the annular member, detect an outer edge of the annular member for each line in the processing target region, and generate an annular member image. The specified annular member image is searched to detect the chamfering position of the annular member, the range corresponding to the chamfering position of the annular member image is searched to detect the contact surface, and the state of the contact surface Based on the above, the quality of the annular member is determined.
[0011]
According to the above configuration, the annular member side to be inspected rotates or revolves the camera side disposed at a predetermined distance from the annular member to capture the outer peripheral portion of the annular member and generate an image signal. The processing target region is determined based on the image signal of the annular member, the outer edge of the annular member is detected for each line in the processing target region, the annular member image is specified, and the specified annular member image Search the inside to detect the chamfering position of the annular member, search the range corresponding to the chamfering position of the annular member image to detect the contact surface, and determine the quality of the annular member based on the state of the contact surface judge.
[0012]
Furthermore, in claim 4, in an image processing method for capturing an image signal of an annular member obtained by an imaging mechanism and processing the image signal, an outer shape of the annular member is detected from the captured image signal, The image processing apparatus is characterized in that after the image of the annular member is corrected to a straight shape, image processing is performed on the image.
[0013]
According to the above configuration, the outer shape of the annular member is detected based on the image signal of the annular member obtained by the imaging mechanism, and the image of the annular member is corrected to a straight shape. Processing is performed, which facilitates image processing of the annular member.
[0014]
Further, according to claim 5, the image signal of the annular member obtained by the imaging mechanism is captured, and the image is processed while changing the threshold value when the image of the annular member is processed. Yes.
[0015]
According to the above configuration, when processing the image of the annular member based on the image signal of the annular member obtained by the imaging mechanism, the image is processed while changing the threshold value, thereby performing image processing. The processing accuracy is improved.
[0016]
Further, in claim 6, when the image signal of the annular member obtained by the imaging mechanism is taken and processing is performed on the image of the annular member, a desired peak of the annular member is combined with a differential peak detection result and a weight function. It is characterized by detecting a part.
[0017]
According to the above configuration, when processing the image of the annular member based on the image signal of the annular member obtained by the imaging mechanism, the desired portion of the annular member is combined with the differential peak detection result and the weight function. Is detected, thereby increasing the detection accuracy of the desired portion.
[0018]
DETAILED DESCRIPTION OF THE INVENTION
"Constitution"
FIG. 1 is a block diagram showing an embodiment of an annular member outer periphery inspection device and an image processing method according to the present invention.
[0019]
The annular member outer periphery inspection apparatus 1 shown in this figure includes an imaging mechanism 2 and a control processing mechanism 3, and the top ring shown in FIG. 2A or the second ring shown in FIG. The piston ring 4 is taken in and rotated, and the outer peripheral portion is imaged, and the image signal obtained by the imaging operation is processed by the control processing mechanism 3 to detect the free gap of the piston ring 4 and the outer edge. Detection, chamfering position detection, contact surface detection, pass / fail determination, display image creation, etc. are performed, as shown in FIG. 3A, whether the position of the contact surface is deviated from the specified range is shown in FIG. Whether or not the width of the contact surface is within the specified range, whether or not the contact surface is cut off in the middle as shown in FIG. 3C, and the like are finally determined. As described above, the outer periphery inspection target of this embodiment includes a top ring having a drum-shaped cross section as shown in FIG. 2A and a top ring having a drum shape as shown in FIG. There is a second ring in which the cross section of the outer peripheral portion is inclined. In the top ring, a contact surface is formed at the center of the outer periphery from the shape, whereas in the case of the second ring, the contact surface is formed at the most protruding portion. In this embodiment, the outer periphery inspection of these contact surfaces is performed by image processing.
[0020]
The imaging mechanism 2 generates various control signals necessary for transporting the piston ring 4 based on the control data DO output from the control processing mechanism 3, and status data DI indicating the transport state of the piston ring 4. A transport control device 5 that generates and supplies this to the control processing mechanism 3 and a piston ring 4 to be inspected from the previous process based on various control signals output from the transport control device 5 and inspects them. The piston ring 4 that is transported to the station device 6 and transports the inspected piston ring 4 to the next process, and the image signal output from the control processing mechanism 3 is captured, and this is inspected by the inspection station device 6 4 is provided with an image monitor device 8 that displays the image as four images.
[0021]
Furthermore, the imaging mechanism 2 generates a light source for illumination light necessary to illuminate the piston ring 4 when power is supplied from the control processing mechanism 3, and the illumination power source device 9 A plurality of light guides 10 that illuminate the outer peripheral surface of the piston ring 4 to be inspected by the illumination light generated in step 1 and a holding mechanism 15 that rotates the piston ring 4 illuminated by each light guide 10 (FIGS. 4 and 5) And an inspection station device 6 including a line sensor 11 that images the outer peripheral surface of the piston ring 4 that is rotated and driven while being held by the holding mechanism 15.
[0022]
Then, while supplying power to each unit based on the commercial voltage “3-phase 200 V” supplied from the commercial power supply, various control signals are sent based on the control data DO output from the control processing mechanism 3 by the transport control device 5. At the same time, based on these various control signals, the transport device 7 is controlled to take in the piston ring 4 to be inspected from the previous process and transport it to the inspection station device 6. After that, the inspection station apparatus 6 receives the illumination light obtained by the illumination power source device 9 by the plurality of light guides 10 provided in the inspection station apparatus 6 and illuminates the outer peripheral surface of the piston ring 4 to be inspected. The piston ring 4 to be inspected is rotated by the holding mechanism 15 provided in the image sensor, the outer peripheral surface of the piston ring 4 is imaged by the line sensor 11, and the image signal obtained thereby is supplied to the control processing mechanism 3. . When the inspection process is completed, the transport control device 5 generates various control signals, and controls the transport device 7 based on the various control signals, so that the piston ring 4 that has finished imaging the outer peripheral surface is Sorted into non-defective and defective products and transported to the next process. In parallel with this operation, an image signal output from the control processing mechanism 3 is captured by the image monitor device 8 and is displayed on the monitor as an image of the piston ring 4 inspected by the inspection station device 6.
[0023]
Further, when the control processing mechanism 3 inspects the piston ring 4, the control processing mechanism 3 generates the control data DO and supplies the control data DO to the imaging mechanism 2 while taking in the situation data DI output from the imaging mechanism 2, and controls this. At the same time, the outer peripheral image (image signal) of the piston ring output from the imaging mechanism 2 is captured to perform free gap detection, outer edge detection, chamfer position detection, contact surface detection, pass / fail determination, display image creation, etc. , An image processing device 11 for determining whether the piston ring 4 is good, an image monitor device 12 for capturing and displaying an image (image signal) of the piston ring 4 obtained by the image processing device 11, a keyboard, a display, A personal computer used as a terminal device or an interface device of the image processing apparatus 11. And Yuta device 13, and a serial printer device 14 for printing out capture print information output from the personal computer apparatus 13.
[0024]
When the piston ring 4 to be inspected is inspected, the image processing apparatus 11 supplies power to each unit based on the commercial voltage “one phase 100 V” supplied from the commercial power supply. The control data DO is generated while capturing the operation contents and the situation data DI output from the imaging mechanism 2 and supplied to the imaging mechanism 2 to control the imaging mechanism 2 and output from the imaging mechanism 2. The outer ring image (image signal) of the piston ring 4 to be inspected is captured, and the piston ring 4 is subjected to free gap detection, outer edge detection, chamfer position detection, contact surface detection, pass / fail determination, display image creation, etc. 4 is judged as good or bad. Further, in parallel with this operation, an image (image signal) of the piston ring 4 to be inspected obtained by the image processing device 11 by the image monitoring device 12 is captured and displayed on the monitor. When print information is output from the image processing apparatus 11, it is captured by the personal computer device 13 and printed out from the serial printer device 14.
[0025]
Next, the operation of this embodiment will be described in detail by dividing it into an overall operation and an inspection operation.
[0026]
<Overall operation>
First, when the inspection of the piston ring 4 to be inspected is performed, the status data DI output from the imaging mechanism 2 is captured by the image processing device 11, and the contents of the status data DI and the operation details of the personal computer device 13 are captured. On the basis of the control data DO, the control data DO is generated and supplied to the transport control device 5 of the imaging mechanism 2.
[0027]
As a result, various control signals are generated based on the control data DO output from the image processing apparatus 11 by the conveyance control device 5, and the conveyance device 7 is controlled. The piston to be inspected by the conveyance device 7 from the previous process. The ring 4 is taken in and transported to the inspection station device 6.
[0028]
A plurality of light guides 10 provided in the inspection station device 6 as shown in FIGS. 4 and 5 while the piston ring 4 sent by the transport device 7 is held by the holding mechanism 15 of the inspection station device 6. The line sensor 11 images the outer peripheral surface of the piston ring 4 while being illuminated, and the image signal (the image signal having the contents shown in FIG. 6) obtained thereby is image processed by the control processing mechanism 3. Supplied to the device 11.
[0029]
As a result, an outer peripheral image (image signal) of the piston ring output from the imaging mechanism 2 is captured by the image processing device 11, and based on this outer peripheral image, a free gap detection process, an outer edge detection process, a chamfer position of the piston ring 4 are performed. Detection processing, contact surface detection processing, quality determination processing, display image creation processing, and the like are performed, and the quality of the piston ring 4 is determined. Then, based on the determination result, control data DO serving as selection information is created, which is supplied to the conveyance control device 5 of the imaging mechanism 2 and the piston ring 4 held by the holding mechanism 15 by the conveyance device 7 is a non-defective product. The defective products are sorted out and only the non-defective products are taken out to the next process. In parallel with this operation, a monitor image (image signal) of the piston ring 4 to be inspected is supplied to the monitor devices 8 and 12 and displayed.
[0030]
Thereafter, the above-described operation is repeated, and the piston ring 4 is sequentially taken from the previous process, and the quality is inspected.
[0031]
When this inspection processing is completed, print information obtained by tabulating the inspection results of the piston rings 4 is output from the image processing device 11, which is captured by the personal computer device 13 and printed out from the serial printer device 14. The
[0032]
<Inspection operation>
Next, according to the flowcharts shown in FIGS. 7 to 9, the details of the above-described free gap detection processing, outer edge detection processing, chamfer position detection processing, contact surface detection processing, pass / fail determination processing, display image creation processing, and the like of the piston ring 4 are described. Explained.
[0033]
<Free gap detection processing (from step ST1 to ST5)>
This free gap detection process can be broadly classified as follows: (1) Processing from image input to projection processing area determination (step ST1), (2) Horizontal projection calculation and horizontal threshold value calculation to detect a free gap (Steps ST2 to ST5). The processing from the image input to the projection processing area determination in (1) further includes the number of effective lines for calculating the number of effective lines in the input screen by detecting the rise and fall of the inspection start signal given from the conveyance control device. The calculation process is divided into a projection process area determination process for obtaining a horizontal projection within a range determined from the input image.
[0034]
(1) Processing from image input to projection processing area determination (effective line number calculation and projection processing area determination)
First, the image processing apparatus 11 captures an outer peripheral image (input image) of the piston ring 4 to be inspected output from the imaging mechanism 2 and also performs an inspection start signal based on the situation data DI output from the transport control device 5. The rise position and fall position of the image are determined, the image to be processed is cut out from the outer peripheral image of the piston ring based on the determination result, and the number of effective lines of the cut out image is calculated based on a predetermined arithmetic expression. To do.
[0035]
That is, the rise time T of the inspection start signal shown in FIG._SAnd falling time T_EIs equivalent to 1.2 rounds of the captured outer peripheral image. The outer periphery image for 1.2 turns is used as an effective line. The number of effective lines is expressed by the following equation.
Effective line number Le = (T_E-T_S) * (64 × 10^-6) / T_Scan
However, T_Scan: Scan time [sec]
[0036]
Next, the image processing apparatus 11 performs the upper clip range Y within the effective line range as shown in FIG._1Clip, Lower clip range Y_2Clip, Left margin range X_{1 LM}, Right margin range X_{2 RM}The processing area from which the is removed is extracted (step ST1).
[0037]
(2) Processing from horizontal projection calculation to free gap detection (horizontal projection calculation and horizontal projection search (free gap detection))
Next, processing for obtaining a horizontal projection within the range of the extracted processing region is performed as follows.
[0038]
In this process, an initial threshold value is determined for each line, the threshold value is changed following the density fluctuation for each pixel, and a density value exceeding this threshold value is added to obtain a horizontal density for one line. Determine the addition value. Here, the threshold value is made to follow the density fluctuation in order to obtain an accurate calculation result by changing the threshold value even for a gentle fluctuation. Then, the horizontal density addition value of all lines is obtained as the horizontal projection value of the processing target area, and the free gap is obtained from the horizontal projection value. More specifically, first, as shown in FIG. 12, one of the lines constituting the processing region is selected and set as a processing target line, and the density cross section of the processing target line is expressed by the following equation. To calculate the initial threshold T_{0 FG}Ask for.
T_{0 FG}= G₀・ Th_HOR+ B_SDC
Where g₀: Start position X of the line to be processed_SConcentration value at
Th_HOR: Preset horizontal projection offset ratio
B_SDC: Edge search binarization minimum value set in advance
[0039]
Thereafter, the image processing apparatus 11 starts the start position X of the processing target line._SThen, a search is performed in the right direction, and a change amount Δg between the density value g (x + Δx) of the next pixel selected in this process and the density value g (x) of the original pixel is calculated as shown in FIG. This amount of change Δg is the edge search binarization minimum value B_SDCIf smaller, the value obtained by adding the change amount Δg to the previous threshold value T is set as a new threshold value T, and Δg is the edge search binarization minimum value B_SDCIf it is not smaller, the update of the threshold value T is skipped. However, by updating the threshold value T, the threshold value T becomes the initial threshold value T._{0 FG}If it is smaller, the initial threshold value T_{0 FG}Is a new threshold T. Thus, the value of the threshold value T varies according to the change amount Δg, but the initial threshold value Tg._{0 FG}Do not get smaller.
[0040]
Thereafter, the image processing apparatus 11 compares the new threshold value T with the density value g (x) of the pixel to be processed, and the density value g (x) of the pixel to be processed becomes the new threshold value T. When larger, the density value g (x) of the pixel to be processed is added to the total density (horizontal projection data S (y)) so far for this processing target line, and the horizontal projection data for this processing target line. Update the value of S (y).
[0041]
Thereafter, the image processing apparatus 11 repeatedly performs the above-described threshold value T update processing and horizontal projection data S (y) update processing until the final position of the processing target line is reached, so that all of the processing target lines are present. The horizontal projection data S (y) for the pixel density value g (x) is calculated.
[0042]
Thereafter, the image processing apparatus 11 sequentially selects the next line in the processing area, performs the above-described processing on all the pixels on each line, and performs horizontal projection data S ( y) is calculated (step ST2).
[0043]
When the horizontal projection data S (y) for all the lines is obtained, the image processing apparatus 11 selects the maximum value and the minimum value of each horizontal projection data S (y) obtained for each line. , These are respectively the horizontal projection maximum values S_{Max H}, Horizontal projection minimum S_{Min H}And a projection search threshold Th necessary for determining the free gap position based on the following equation:_HIs calculated (step ST3).
Th_H= S_{Min H}+ (S_{Max H}-S_{Min H}) ・ Th_HR
However, Th_HR: Preset horizontal projection threshold value ratio
[0044]
Thereafter, the image processing apparatus 11 performs the projection search threshold Th as shown in FIG._HAre compared with the horizontal projection data S (y) for each line, and the projection search threshold value Th of each horizontal projection data S (y) is compared._HA line having a smaller value is determined as a free gap portion (step ST4).
[0045]
If the free gap portion is in the middle of the processing region as shown in FIG. 15, the image processing apparatus 11 removes the exclusion target region from the upper end of the free gap portion and removes the length L from the upper processing region.₁And the length L from the lower processing area excluding the exclusion target area from the lower end of the free gap portion.₂These areas are extracted, and these are used as subsequent process target areas (step ST5).
[0046]
However, in this case, these lengths L₁, L₂Is the total outer circumferential length π · D / Y of the piston ring 4 to be inspected._ScaleTo the length excluding the exclusion target portion on the upper end side and the exclusion target portion on the lower end side of the free gap portion.
[0047]
Also, as shown in FIG. 16, if the free gap portion is at the beginning or the end of the processing target area, the image processing apparatus 11 excludes the exclusion target area from each part in contact with the free gap part, and the length L₁+ L₂This area is extracted and set as a subsequent process target area. In this way, the portion excluding the free gap portion and the non-inspection region near the free gap is the inspection target region.
[0048]
<Outer edge detection processing (from step ST6 to ST9)>
(1) Threshold determination
Next, in order to detect the outer edges (referred to as the contour of the piston ring) on the left and right of the piston ring 4 as shown in FIG. Left side of (start position X_S) Density value g (x) is the maximum leftmost density value g_GLL(Global.Left.Level) and the right part of each line (start position X_STo width W_SEThe maximum density value g (x) at the right edge)_GRL(Global.Right.Level). Furthermore, for each line, start position X_STo width W_SEThe maximum density value g of each line_ML(Max.Level) is calculated and each of these maximum density values g_MLOf these, the smallest is the maximum density value g_GML(Global.Max.Level) (step ST6), these maximum density values g_GML, Left end density value g_GLL, Right end density value g_GRLBased on the following formula, the left side threshold Th_LB(Left.Boundary) and the right threshold Th_RB(Right.Boundary)
Th_LB= Th_BR・ (G_GML-G_GLL)
Th_RB= Th_BR・ (G_GML-G_GRL)
However, Th_BR: Predetermined ratio for outer edge search threshold determination
[0049]
(2) Edge search
<Left edge position detection>
Thereafter, the image processing apparatus 11 performs the calculation shown in the following formula to determine the left edge position X of the piston ring 4 from the left side._LELeft side threshold T required to detect_{0 LB}Calculate
T_{0 LB}= Th_LB+ B_SDC
However, B_SDC: Edge search binarization minimum value set in advance
[0050]
Next, the image processing apparatus 11 selects one of the lines in the processing target area, sets this as the processing target line, and then starts the start position X of the processing target line as shown in FIG._STo width W_SEThe amount of change Δg between the density value g (x + Δx) of the next pixel selected in this process and the density value g (x) of the original pixel is calculated while searching rightward up to the point. The absolute value of the change amount Δg is the edge search binarized minimum value B._SDCIf smaller, the value obtained by adding the change amount Δg to the previous threshold value T is set as a new threshold value T, and the absolute value of the change amount Δg is the edge search binarization minimum value B_SDCIf it is larger, the update of the threshold value T is skipped. In this way, the threshold value is changed following the gentle density change.
[0051]
Then, the image processing apparatus 11 compares the new threshold value T with the density value g (x) of the pixel to be processed, and the density value g (x) of the pixel to be processed becomes the new threshold value T. When the pixel position is smaller, the left edge position X of the piston ring 4 with respect to the processing target line_LELeft edge reference point X required to detect_LSIf not, the pixel on the right side is further selected.
[0052]
Thereafter, the image processing apparatus 11 repeats the above-described calculation process of the threshold value T and the comparison process between the pixel density value g (x) and the threshold value T.
[0053]
When the density value g (x) of the pixel selected as the processing target is greater than or equal to the new threshold value T, the image processing apparatus 11 detects the left edge of the piston ring 4 with respect to the position x of this pixel. Point (left edge reference point) X_LSAnd the left edge reference point X as shown in FIG._LSA constant width D set in advance around_BSOnly the density change Δg of each adjacent pixel on the left side on this processing target line and the density change Δg ′ of each adjacent pixel on the right side are respectively calculated, and these density change amounts Δg and Δg ′. Among these, the density change (Δg ′ in FIG. 20) having the largest value is detected, and the position x of the pixel having the maximum density change Δg ′ is detected on the piston ring 4 with respect to the processing target line. Left edge position X_LEIs determined.
[0054]
<Right edge detection>
Next, the image processing apparatus 11 performs the left edge position X described above._LEThe threshold value Th on the right side is_RBThe right edge position X on this processing target line_RERight threshold T necessary to detect_{0 RB}As shown in FIG. 21, the end position of the processing target line (start position X_STo width W_SESearch from left to right) and based on the change amount Δg between the density value g (x−Δx) of the next pixel selected in this process and the density value g (x) of the original pixel. While updating the threshold value T, the right edge reference point X of the piston ring 4_RSWhen this is detected, the right edge reference point X_RSA constant width D set in advance around_BSOnly the density change Δg of each adjacent pixel on the right side on the processing target line and the density change Δg of each adjacent pixel on the left side are respectively calculated, and among these density change Δg, A density change Δg giving a large value is detected, and the position x of the pixel having the maximum density change Δg is determined as the right edge position X of the piston ring 4 with respect to the processing target line._REIs determined.
[0055]
Thereafter, the image processing apparatus 11 repeatedly performs the above-described processing while sequentially updating the lines to be processed, and the left edge position X of the piston ring 4 for each line._LEAnd right edge position X_REAre detected (steps ST7 and ST8).
[0056]
(3) Determination of processing area
And the left edge position X of the piston ring 4 with respect to all lines_LEAnd right edge position X_REAs shown in FIG. 22, the image processing apparatus 11 detects the left edge position X for each line._LEAnd right edge position X_REMidpoint (center position X of piston ring 4_C) And the width W of the piston ring 4_RingCalculate After that, the leftmost center position X_CLAnd the rightmost center position X_CRAnd select this as the center position X_CAnd the ring width W_RingMinimum value W_{R Min}And the maximum value W_{R Max}And select the ring width W_RingThe wobble data.
[0057]
Thereafter, the image processing apparatus 11 is configured so that each central position X of the piston ring 4_CThe maximum value W of the ring width (center line) and the ring width_{R Max}And the extraction area width W on the left side of the center line of the piston ring 4 is calculated._Lp, Right extraction area width W_RpThe left extracted region width W from the processing target region_Lp, Right extraction area width W_RpIs cut out, and this is used as a subsequent processing target area.
W_Lp= (W_{R Max}/ 2) ・ X_{1 LC}
W_Rp= (W_{R Max}/ 2) ・ X_{2 RC}
However, X_{1 LC}: Preset left margin for processing image display
X_{2 RC}: Preset right margin for processing image display
[0058]
Thereafter, the image processing apparatus 11 shifts each line in the lateral direction so that the center line of the piston ring 4 is a straight line, and extends the outer shape of the piston ring 4 straight (step ST9).
[0059]
<Chamfering position detection processing (steps ST10 to ST13)>
(1) Vertical projection calculation
Next, as shown in FIG. 23, the image processing apparatus 11 divides the image of the piston ring 4 in the processing target area into 16 parts in the vertical direction, selects one of each of the divided areas, and divides this into the processing target split parts After forming the region, as shown in FIG. 24, the density data g (x) of the pixels in each line in the processing target divided region is added in the vertical direction to obtain the vertical projection data f (x) (FIG. 8). Step ST10).
[0060]
(2) Chamfer position search
(1) Threshold determination for projective data search
Next, the left side chamfering position X_LCAnd right chamfering position X_RCLeft-side initial threshold Th necessary for searching for_LPAnd right initial threshold Th_RPAnd ask.
Th_LP= Th_LB・ L
Th_RP= Th_RB・ L
However, Th_LB: Left threshold value obtained by outer edge detection processing
Th_RB: Right threshold value obtained by outer edge detection processing
L: Number of lines constituting the processing target divided area
[0061]
(2) Projection search (edge detection)
Thereafter, the image processing apparatus 11 performs the calculation shown in the following equation, and the leftmost end position X with respect to the processing target divided region._LEChamfering position X on the left side of the piston ring 4 from the side_LCLeft side threshold T required to detect_{0 LC}Calculate
T_{0 LC}= Th_LP+ F (x_LE)
Where f (x_LE): Leftmost position x of the vertical projection data f (x) of the processing target divided area_LEThe value of the
[0062]
Thereafter, the image processing apparatus 11 updates the threshold value T (x) based on the following equation:
[Expression 1]

Where t₁: T₁= B_SDC.Threshold value represented by L
As shown in FIG. 25, the leftmost position x of the processing target divided area_LEAnd the right projection data f (x) at the next position x selected by this search process is compared with the threshold value T (x), and the vertical projection data at the position x to be processed is compared. When f (x) is smaller than the threshold value T (x), this position x is the left chamfering position X of the piston ring 4 with respect to the processing target divided region._LCLeft chamfer reference point X required to detect_LSSThe position x on the right side is further selected, and the calculation process of the threshold value T (x) and the comparison process of the vertical projection data f (x) and the threshold value T (x) are repeated. .
[0063]
When the vertical projection data f (x) at the position x selected as the processing target is greater than or equal to the threshold value T (x), the image processing apparatus 11 determines that the position x is a piston in the processing target divided area. Chamfering position X on the left side of ring 4_LCDetection point (left chamfer position X_LCLeft chamfer reference point when searching for) X_LSSAnd the left chamfer reference point X_LSSAt a point further to the right side from the point where the differential function f ′ (x) of the vertical projection data f (x) satisfies the following formula (left-side highest peak position X_LMP) Is detected.
f ′ (x) ≦ 0
[0064]
Next, the image processing apparatus 11 performs the above-described left maximum peak position X described above._LMPAs shown in FIG. 26, the right-side initial threshold Th is_RPThe right chamfering reference point X in this processing target divided area_RSSRight initial threshold T required to detect_{0 RC}Calculate
[0065]
Thereafter, the image processing apparatus 11 reads the rightmost position X of the processing target divided area._REFrom the left to the right, while updating the threshold value T (x), the vertical projection data f (x) of the next position x selected in this process and the threshold value T (x) Based on the right chamfering position X of the piston ring 4_RCDetection point (right chamfer position X_RCRight chamfer reference point when searching for) X_RSSAnd the right chamfering reference point X_RSSThe point at which the differential function f ′ (x) of the vertical projection data f (x) becomes zero or less at the position x on the left side of_RMP) Is detected.
[0066]
(3) Projection data smoothing and (4) Minimum density position detection
Next, as shown in FIG. 27, the image processing apparatus 11 averages the vertical projection data f (x) of the processing target divided region with a preset moving average width, for example, 2m + 1 width, and this vertical projection data. The smoothing of f (x) and the left-side highest peak position X of the vertical projection data f (x) of the divided area as shown in FIG._LMPAnd the highest peak position on the right side X_RMPAnd range W_PTThe vertical projection data f (x) inside is extracted, and the position where the value of the vertical projection data f (x) is minimum (minimum density position X_{Min L}) Is detected.
[0067]
(5) Search range determined from piston ring standard dimensions
Thereafter, the image processing apparatus 11 performs the calculation shown in the following equation, and based on the standard of the piston ring 4, the left chamfering reference point X as shown in FIG._LSSAnd right chamfer reference point X_RSSChamfer position search width W based on_PredBAsk for.
W_PredB= W_PB・ Max (W_UR, W_LR) / W_Ring
However, W_PB: Left chamfer reference point X_LSSAnd right chamfer reference point X_RSSDistance to
W_UR: Piston ring 4 top side width
W_LR: Piston ring 4 bottom side width
W_Ring: Piston ring 4 width
[0068]
(6) Search range correction
Thereafter, the image processing apparatus 11 performs the left chamfering reference point X as shown in FIGS._LSSTo the minimum density position X of the vertical projection data f (x)_{Min L}Distance and chamfer position search width W_PredBThe smaller one of these values is the left chamfer position X_LCAnd the right chamfering reference point X_RSSTo the minimum density position X of the vertical projection data f (x)_{Min L}Distance and chamfer position search width W_PredBThe smaller one of these values is the right chamfer position X_RCSearch range.
[0069]
(7) Differentiation of projection data within the search range
Next, the image processing apparatus 11 performs the left chamfering position X._LCLeft peak position X within the search range_LMPWhile changing the position x from the inside to the inside, the calculation shown in the following equation is performed to obtain the differential value f ′ (x) for each of these positions x.
f ′ (x) = {f (x + D_IN) -F (x)} / D_IN
However, D_IN: Preset differential interval value
[0070]
(8) Search for differential value f '(x)
If the peak of these differential values f ′ (x) is detected and satisfies the following equation, the position x is changed to the left chamfering position X._LCAnd
| F ′ (x) | ≧ Th_PS
However, Th_PS: Preset threshold value for chamfer position detection
In addition, when the above condition is satisfied, the calculation shown in the following formula is performed and the threshold value Th_PSTo change.
Th_PS= | F '(x) | .Th_PSR
However, Th_PSR: Preset chamfer position detection peak detection rate
[0071]
Thereafter, the image processing apparatus 11 uses the left chamfering position X._LCThe right chamfered position X_RCIs detected.
[0072]
Hereinafter, the image processing apparatus 11 calculates the vertical projection data f (x) for each remaining divided region, and the left and right chamfering reference points X._LSS, X_RSSSearch process, left and right peak peak position X_LMP, X_RMPDetection processing, smoothing processing of vertical projection data f (x), minimum density position x_{Min L}Detection processing, search range determination processing, left and right chamfer positions X in the search range_LC, X_RCThe left and right chamfer positions X of each divided area_LC, X_RCIs obtained (steps ST11 and ST12).
[0073]
▲ 9 ▼ Chamfer position correction
The chamfering position obtained from the projection data of the 16-divided image is corrected by performing the following processing.
[0074]
As shown in FIG. 32, the image processing apparatus 11 performs the left and right chamfering positions X of each divided area._LC, X_RCLeft side, right side average value X_ALC, X_ARCAnd left and right side standard deviation σ_LC, Σ_RCAre calculated, and then the left and right chamfer positions X of each divided region are calculated._LC, X_RCFor example, left chamfering position X_LCIs left average X_ALCTo calculate whether it falls within the range of values shown in the following formula,
σ_LC・ B_PRE
However, B_PRE: Pre-set chamfer position appropriate evaluation coefficient
Each left chamfer position X_LCLeft chamfering position X not within this range_LCIf there is, this left chamfering position X_LCThe left chamfering position X of the divided region that is determined to have a large variation as shown in FIG._LC, Each left chamfer position X of the divided area before and after_LCIs linearly interpolated as indicated by the dotted line.
[0075]
In addition, the image processing apparatus 11 determines the left chamfer position X of each divided area._LCThe right chamfering position X of each divided region in the same procedure as the variation determination processing of_RCIs the right average X_ARCTo calculate whether it falls within the range of values shown in the following formula,
σ_RC・ B_PRE
However, B_PRE: Pre-set chamfer position appropriate evaluation coefficient
Each chamfer position X_RCOf these, the data not within this range is judged to have a large variation, and the right chamfering position X of the divided area determined to have a large variation._RC, Each right chamfering position X of the divided areas before and after_RCBased on the above, linear interpolation is performed (step ST13).
[0076]
<Contact surface detection processing (from step ST14 to ST20)>
(1) Hit position search range
As shown in FIG. 34, the image processing apparatus 11 performs the calculation shown in the following formula for each of the 16 divided areas, and calculates the left and right chamfer positions X._LC, X_RCSearch inside from left side, right side hit position X_LE, X_RENon-inspection area width W required when detecting_NSBAfter asking
W_NSB= Th_SBR・ (X_RC-X_LC)
However, Th_SBR: Hit position non-inspection area ratio
X_RC: Right chamfer position
X_LC: Left chamfer position
Left side position X_LCRight chamfering position X_RCThe left chamfer position X of the range up to_LCNon-inspection area width W inside_NSBRight side chamfering position X_RCNon-inspection area width W inside_NSBThe range excluding the left side, right side hit position X_LE, X_REThe search range is necessary to detect.
[0077]
(2) Projection data classification
Thereafter, as shown in FIG. 35, the image processing apparatus 11 divides the search range into a plurality of parts so that each of the 16 divided areas has a small section from a peak to a peak in the search range. For each of these subsections, the start position of the section, the end position of the section, the position of the minimum value in the section are obtained, and the position of the minimum value and the section number in all sections, and the value at the position of the minimum value (minimum value) Ask for.
[0078]
(3) Linking small sections between adjacent screens
After that, as shown in FIG. 36, each subsection of each divided area is connected, and the correct left side and right side hit position X_LE, X_REIn order to find out, the following processing is started.
[0079]
First, the image processing apparatus 11 divides each sub-section of the sub-region (processing target sub-region) to be processed, for example, the processing target sub-region into first, second, and third sub-sections as shown in FIG. If so, position (minimum position) X giving the minimum value in these first subsections_{1 Min}And the position that gives the minimum value in the second subsection (minimum position) X_{2 Min}And position (minimum position) X giving the minimum value in the third subsection_{3 Min}And detect.
[0080]
Then, the image processing apparatus 11 divides each small section of the divided area (lower adjacent divided area) below the processing target divided area, for example, the lower adjacent divided area into first and second small sections as shown in FIG. If so, as shown in FIG. 38, the first, second, and third small sections of the processing target divided area correspond to the first and second small sections of the lower adjacent divided area, and the lower adjacent divided area Minimum position X of the processing target divided region of the vertical projection data f (x)_{1 Min}, X_{2 Min}, X_{3 Min}Value f (X_{1 Min}), F (X_{2 Min}), F (X_{3 Min}) Is created with the evaluation values of the first and second subsections of the lower adjacent subregion for the first, second and third subsections of the processing target subregion.
[0081]
Next, the image processing apparatus 11 performs the above-described processing on each small section of the divided area (upper adjacent divided area) that is adjacent to the processing target divided area, so that the first, second, Evaluation value f (X of each small section of the upper adjacent divided region with respect to the third small section_{1 Min}), F (X_{2 Min}), F (X_{3 Min}) To create a join table.
[0082]
Thereafter, the image processing apparatus 11 performs the same calculation for the remaining divided areas, and creates a combined table in which each divided area is a processing target divided area.
[0083]
Thereafter, the image processing apparatus 11 evaluates each evaluation value f (X_{1 Min}) To f (x_{n Min}), And, among these total values, the smallest subsection of the total value, that is, the correct left side / right side position X_LE, X_REAre selected, and these subsections are connected (step ST14).
[0084]
(4) Creation of weight function
Next, the image processing apparatus 11 selects one of the divided areas, and designates this as a processing target divided area as the left and right hit positions X._LE, X_REAs shown in FIG. 39, the maximum value and the minimum value of the vertical projection data f (x) within the search range are detected as well as the data necessary for creating the weighting function used in the detection processing of The density value at the maximum position (maximum density value) g by dividing the minimum value by the number of lines in the processing target divided area_{P Max L}And the density value at the minimum position (minimum density value) g_{P Min L}And the maximum concentration value g_{P Max L}And the minimum density value g_{P Min L}Based on the above, the main concentration level M_mAsk for.
M_m= G_{P Min L}+ (G_{P Max L}-G_{P Min L}) ・ B_MPR
However, B_MPR: Preset weight function coefficient rate
[0085]
Thereafter, the image processing apparatus 11 uses the main density level M._mAnd the maximum density value g_{P Max L}And the minimum density value g_{P Min L}And each preset weight function coefficient W_B(Bottom / Weight), W_TBased on (Top · Weight), a weight table of the weight function M having the shape shown in FIG. 40 is created.
[0086]
(5) Correction of search range
Next, if the piston ring 4 to be inspected is a top ring, the image processing apparatus 11 leaves the search range as shown in FIG. 41, and if the piston ring 4 to be inspected is a second ring. 42, a hit position search excess area A set in advance as shown in FIG._ESOnly the right chamfer position X_RCExtend the search range to the right.
[0087]
Thereafter, the image processing apparatus 11 performs the same calculation for the remaining divided areas, creates a weight table for each divided area, and determines a search range (steps ST15 and ST16).
[0088]
(6) Differentiation of concentration data g (x)
After that, the image processing apparatus 11 selects a divided area to be processed from among the divided areas, performs the calculation represented by the following formula for each line in the processed divided area, and performs density data g of the piston ring 4. A differential value g ′ (x) of (x) is obtained.
g ′ (x) = {g (x + D_IN) -G (x)} / D_IN
However, D_IN: Differential interval value used in chamfer position detection processing
[0089]
(7) Winning position
Next, as shown in FIG. 43, the image processing apparatus 11, among the small sections constituting the processing target divided area, the mainstream left side and right side hit positions X obtained by the above-described small section connection processing._LE, X_REAnd a position indicating the minimum density in the second small section of the density data g (x) of the line to be processed. (Minimum concentration position X_Min) Is detected.
[0090]
Thereafter, the image processing apparatus 11 reads the minimum density position X._MinThe maximum density value g within the search range using the density value (minimum density value) at_{Max L}And the minimum density value g_{Min L}And the maximum evaluation value E_MaxAnd position X giving the maximum evaluation value_{Max E}After initializing the minimum density position X_MinThe left search range is searched while updating the value of the position x starting from.
[0091]
In this search process, the image processing apparatus 11 determines that the density data g (x) for the position x is the maximum density value g._{Max L}If it is smaller, the density data g (x) is the minimum density value g._{Min L}Whether the density data g (x) is the minimum density value g_{Min L}If it is smaller, the minimum density value g_{Min L}Update the value of.
g_{Min L}= G (x)
[0092]
In this search process, the density data g (x) for the position x is the maximum density value g._{Max L}If larger, the image processing apparatus 11 checks whether or not the differential value g ′ (x) at the position x is a peak, and if the differential value g ′ (x) is a peak, the differential value g ′ (x). And the weighting value M corresponding to the value of the density data g (x) at the position x with reference to the weighting table described above, the calculation shown in the following equation is performed, and the evaluation function E_VSeeking
E_V= G '(x) · M
Furthermore, calculate whether the following equation is satisfied,
E_V・ C_EV> E_Max
However, C_EV: Damping coefficient of evaluation function
If this is satisfied, the maximum evaluation value E_MaxAnd position X giving the maximum evaluation value_{Max E}Update.
E_Max= E_V
X_{Max E}= X
[0093]
And the minimum density position X_MinTo the left end of the left search range, the image processing apparatus 11 obtains the maximum evaluation value E obtained at this time._MaxAnd the maximum evaluation value E_MaxPosition X giving_{Max E}And the left search range hit position (left hit position) X_LEIs determined.
[0094]
After that, the image processing apparatus 11 determines the hit position X of the left search range described above._LEThe minimum concentration position X_MinThe maximum density value g in the search range using the density value (minimum density) at_{Max L}And the minimum density value g_{Min L}And the maximum evaluation value E_MaxAnd position X giving the maximum evaluation value_{Max E}, The minimum density position X is updated while updating the value of the position x._MinThe right search range is searched from the right and the right search range hit position (right hit position) X_RECalculate
[0095]
Hereinafter, the image processing apparatus 11 determines the left side hit position X for each remaining line in the processing target divided area._LEAnd right side hit position X_REAnd calculate.
[0096]
And the left side hit position X for each line in the processing target divided area_LEAnd right side hit position X_REWhen the calculation is completed, the image processing apparatus 11 sequentially selects the remaining divided areas and sets them as processing target divided areas, and then performs the above-described processing to perform the left hit position X for each line._LEAnd right side hit position X_REAre calculated (step ST17).
[0097]
(8) Correction of hit position (fixed level search)
Next, the image processing apparatus 11 selects a line to be processed among all the lines as shown in FIG. 44 in order to correct a jump due to dirt or the like, and a left side hitting position X with respect to this line (target line)._LEConcentration data g (x) with respect to the left and right positions X of the two lines before and after the target line_LEThe average value (left average density data) T of the three density data g (x) excluding the maximum density data g (x) and the minimum density data g (x) based on the density data g (x)_LACalculate
[0098]
Thereafter, the image processing apparatus 11 reads the left average density data T_LAChecks whether or not
T_LA<B_{S Min L}+ G_{L Min L}
However, B_{S Min L}: Preset binary search minimum value
g_{L Min L}: Minimum density value of the small section including the target line
If this condition is satisfied, the left-side average density data T based on the following formula:_LATo prevent it from becoming too small.
T_LA= B_{S Min L}+ G_{L Min L}
[0099]
Thereafter, the image processing apparatus 11 performs the left average density data T on the target line._LAThe position x of the density data g (x) having a larger value is detected, and this position x is detected as the correct left side hitting position X._LELeft side hit position X on the target line_LETo correct.
[0100]
Thereafter, the image processing apparatus 11 sequentially selects the next line among the remaining lines and makes this the target line, while the left hit position X on the target line described above._LEMake corrections.
[0101]
Next, the image processing apparatus 11 detects the right side hit position X on all the lines._REFor the left side hit position X described above_LEPerform the same process as the correction process of_REIs corrected (step ST18).
[0102]
(9) Correction of hit position (jump correction)
After that, the image processing apparatus 11 selects one of the lines in the entire divided area and makes it a target line, and then the left side hit position X of the target line._LEBefore and after this line of interest_LEAnd when these differences are positive, that is, as shown in FIG._LEThe left side hit position X of the line of interest_LEWhen is outside, check whether the following equation is satisfied,
a_1b> Limit_JP
a_2a> Limit_JP
However, Limit_JP: Pre-set hit position jump limit value (outside)
a_1b: The hit position X of the line before the attention line_LEAnd the hit position X of the line of interest_LEDifference from
a_2a: Hit position X of the line after the attention line X_LEAnd the hit position X of the line of interest_LEDifference from
When these are satisfied at the same time, the left side hit position X of the front and rear lines_LEThe left side hit position X of the line of interest_LEEach left side hit position X on the front and back line_LELeft side hit position X on the line of interest based on_LEThis is corrected by linear interpolation.
[0103]
In addition, as shown in FIG._LEThe left side hit position X of the line of interest_LEIs on the inside, the image processing apparatus 11 checks whether or not the following equation is satisfied,
b_1b> Limit_JM
b_2a> Limit_JM
However, Limit_JM: Pre-set hit position jump limit value (inside)
b_1b: The hit position X of the line before the attention line_LEAnd the hit position X of the line of interest_LEDifference from
b_2a: Hit position X of the line after the attention line X_LEAnd the hit position X of the line of interest_LEDifference from
When these are satisfied at the same time, the left side hit position X of the front and rear lines_LEBased on the left side hit position X of the line of interest_LEThis is corrected by linear interpolation.
[0104]
Thereafter, the image processing apparatus 11 sequentially selects the next line among the remaining lines and makes this the target line, while the left hit position X on the target line described above._LEThe position X of each left side is corrected._LESmooth the line connecting.
[0105]
Next, the image processing apparatus 11 receives the left side hit position X described above._LEThe right hit position X of each line in all the divided areas in the same procedure as that used for the correction processing of_REThis is corrected by the linear interpolation method and the right side hit position X_REIs smoothed (step ST19).
[0106]
<Spot position smoothing>
Next, a portion where a large jump is found from the hit position data is searched, and smoothing is performed. The image processing apparatus 11 selects one of the lines constituting all the divided areas and sets this as the attention line, and then, as shown in FIG._LEAnd a certain number of lines (jump position detection interval M_{1 IN}) A left-side hitting position X of a line (a reference line) that is separated downward by a)_LEThe difference Δb is calculated, and it is determined whether or not the difference Δb is larger than the difference Δb.
[0107]
When the attention line difference Δb is larger than the difference Δb so far, the image processing apparatus 11 determines that the number of lines (jump position detection interval M) from the attention line as shown in FIG._{2 IN}) A left-side hit position X of a line (a re-reference line) that is separated downward by a)_LEAnd a fixed number of lines from the reference line (jump position detection interval M_{2 IN}) A left-side hit position X of a line (a re-reference line) that is separated upward by_LEThe difference Δw from is calculated.
[0108]
When the difference Δw is larger than the maximum value of the difference Δw obtained by the processing so far, the image processing apparatus 11 stores the difference Δb as the maximum value (maximum difference) Max (Δb) so far.
[0109]
Thereafter, the image processing apparatus 11 performs the above-described calculation process of the differences Δb and Δw for the remaining lines while shifting the target line by line, and the newly obtained difference Δb is the maximum of the difference Δb so far. When larger than the value, it is determined whether or not the newly obtained difference Δw is larger than the previous difference Δw, and when the newly obtained difference Δw is larger than the previous difference Δw, the newly obtained difference The maximum difference Max (Δb) is updated with Δb.
[0110]
When the maximum difference Max (Δb) for all lines is obtained, the image processing apparatus 11 calculates whether or not the maximum difference Max (Δb) satisfies the following expression.
Max (Δb) ≧ L_JS
However, L_JS: Minimum preset jump length
[0111]
If this condition is satisfied, the image processing apparatus 11 detects the left-side hit position X including the line having the maximum difference Max (Δb) among the lines as shown in FIG._LEIs determined as a jump point, the calculation shown in the following formula is performed, and each left side hit position X including this jump point is calculated._LESmoothing length L required to smooth the jump part of the line connecting_SmAnd the smoothing length L from the jump point_SmThe starting point Y of the smoothing process just above the line_SAnd the smoothing length L from the reference point corresponding to the jump point._SmOnly the line below the end point Y of the smoothing process_EnTo decide.
L_Sm= Max (Δb)
・ {C_{Sm F}・ D / (B-b_Up-B_Lo)}
However, C_{Sm F}: Preset smooth length calculation coefficient
D: Diameter of the piston ring 4 to be inspected (mm)
B: Thickness (mm) of the piston ring 4 to be inspected
b_Up: Top surface removal of the piston ring 4 to be inspected (mm)
b_Lo: Bottom surface removal of piston ring 4 to be inspected (mm)
[0112]
Thereafter, the image processing apparatus 11 starts the start point Y as shown in FIG._SSmoothing length M set in advance from_{3 Sm}Is determined as the upper boundary processing section, and the left side hit position X of each line in the upper boundary processing section is determined._LEAnd a function f (y), and the end point Y_EnSmoothing length M set in advance from_{3 Sm}Is determined to be the lower boundary processing section, and the left side hit position X of each line in the lower boundary processing section is determined._LEIs a function f (y), and the left side hit position X of each line in these upper boundary processing sections_LEAnd the hit position X of each line in the lower boundary processing section_LEA straight line approximation by the least square method is performed on the function f (y) indicating
[0113]
Next, the image processing apparatus 11 creates a cubic curve that smoothly connects the lower end of the approximate straight line in the upper boundary processing section and the upper end of the approximate straight line in the lower boundary processing section using a smooth function as shown in FIG. And the left side hit position X of each line_LESmooth the jump part of the line segment that connects_LEMake the whole line connecting
[0114]
And left side hit position X_LEWhen the smoothing process is completed, the image processing apparatus 11 determines that the left side hit position X described above._LEIn the same procedure as the smoothing process of_RETo the right side hit position X_RESmooth the jump part of the line connecting_REAre continuous (step ST20).
[0115]
<Determination process (from step ST21 to ST23)>
(1) Inspection of contact width
The image processing apparatus 11 determines the left side hit position X for each line._LEX and the right contact position X_REThe distance (per contact width) W from the line connecting₁As shown in FIG. 52, these contact widths W are calculated.₁Among them, a predetermined length L from the joint portion (free gap portion) of the piston ring 4_TPer unit width W₁Calculates whether the following condition is satisfied.
W₁≧ Limit_LoBT
W₁≦ Limit_UpBT
However, Limit_LoBT: Lower limit of hit width at joint
Limit_UpBT: Upper limit of hit width of joint
[0116]
Thereafter, the image processing apparatus 11 has a predetermined length L from the joint portion._TThe contact width W of the part excluding the part of₁Calculates whether the following condition is satisfied.
W₁≧ Limit_LoB
W₁≦ Limit_UpB
However, Limit_LoB: Lower limit of hit width
Limit_UpB: Upper limit of hit width
[0117]
If these conditions are satisfied at the same time, the image processing apparatus 11 has a contact width W of the piston ring 4.₁If at least one of these conditions is not satisfied, the contact width W of the piston ring 4 is determined.₁Is determined to be defective (step ST21 in FIG. 9).
[0118]
(2) Inspection of remaining width
Next, if the piston ring 4 is a top ring, the image processing apparatus 11 determines the left chamfering position (upper side chamfering position) X for each line as shown in FIG._LCAnd the left side contact position X of the contact surface_LEDistance (upper remaining width) W_{2 UpRe}And right chamfering position (lower chamfering position) X_RCAnd the right side contact position X of the contact surface_RCDistance (lower remaining width) W_{2 LoRe}And the upper remaining width W_{2 UpRe}And lower remaining width W_{2 LoRe}And whether or not the following equation is satisfied.
W_{2 UpRe}≦ Limit_WURT
W_{2 LoRe}≧ Limit_WLRT
However, Limit_WURT: Upper limit of upper remaining width for top ring
Limit_WLRT: Lower left side lower limit for top ring
[0119]
If these conditions are satisfied at the same time, the image processing apparatus 11 determines that the upper remaining width W of the piston ring._{2 UpRe}And lower remaining width W_{2 LoRe}If at least one of these conditions is not satisfied, the upper remaining width W is determined._{2 UpRe}The lower remaining width W_{2 LoRe}The lower remaining width W_{2 LoRe}Upper remaining width W_{2 UpRe}The upper remaining width W described above_{2 UpRe}Pass / fail judgment processing, lower remaining width W_{2 LoRe}If the conditions are satisfied by these pass / fail determination processes, the piston ring 4 is turned upside down and the remaining width is determined as a non-defective product.
[0120]
In addition, even if such processing is performed, if the above-described conditions are not satisfied at the same time, the image processing apparatus 11 determines the upper remaining width W of the piston ring 4._{2 UpRe}Or lower remaining width W_{2 LoRe}It is determined that at least one of these is defective.
[0121]
Further, if the piston ring 4 is a second ring, the image processing apparatus 11 is provided with the lower side chamfering position X for each line as shown in FIG._RCAnd the right side contact position X of the contact surface_REDistance (lower remaining width) W_{2 LoRe}And the lower remaining width W_{2 LoRe}Computes whether or not
W_{2 LoRe}≦ Limit_WURS
However, Limit_WURS: Lower remaining upper limit for the second ring
[0122]
If this condition is satisfied, the image processing apparatus 11 determines that the piston ring 4 has a lower remaining width W._{2 LoRe}If this condition is not satisfied, the lower remaining width W of the piston ring 4 is determined._{2 LoRe}Is determined to be defective (step ST22).
[0123]
(3) Winning
Next, the image processing apparatus 11 performs the above-described left and right contact positions X for each line._LE, X_RELeft hit position X obtained by the detection process_LEEvaluation function E_VAnd right side hit position X_REEvaluation function E_VAverage value E_AVAnd a threshold value Th necessary for determination of hitting (referring to a portion where the hit is broken. The hit portion is brighter than the surroundings) based on the following equation.₂Ask for.
Th₂= E_AV・ E_{V LoRa}
However, E_{V LoRa}： Contrast lower limit ratio set by experience
[0124]
Thereafter, the image processing apparatus 11 performs the above-described left and right contact positions X for each line._LE, X_REConcentration value g obtained by the detection process of_{Min L}Among these, the smallest density value g of the smallest value_{Min L}And select the minimum concentration value g_{Min ML}As shown in the following equation, the threshold value Th required for determining hit_ThreeAfter asking
Th_Three= G_{Min ML}・ L_{UMin R}
However, L_{UMin R}: Preset minimum value upper limit ratio
For each line, the minimum density value g_{Min L}Computes whether or not
g_{Min L}> Th_Three
[0125]
Thereafter, the image processing apparatus 11 performs the left side hit position X for each line._LEEvaluation function E_VCalculate whether or not the condition shown in the following equation is satisfied,
E_V<Limit_EVLOr E_V<Th₂
However, Limit_EVL: Lower limit value of the preset evaluation function
Right side hit position X_REEvaluation function E_VCalculates whether the condition shown in the following equation is satisfied.
E_V<Limit_EVLOr E_V<Th₂
[0126]
When all of these conditions are not satisfied, the image processing apparatus 11 determines that the piston ring 4 is not hit, and when one of the conditions is true, the image processing apparatus 11 is hit with the piston ring 4. Judge that there is.
[0127]
(4) Hit center position
Next, as shown in FIG. 55, the image processing apparatus 11 performs the left side hit position X for each line._LEAnd right side hit position X_RETo obtain the center position of the contact surface, as shown in FIG._LEAnd the distance L from the center position of the contact surface_CAnd this distance L_CComputes whether or not
Limit_{Lo C}<L_C<Limit_{Up C}
However, Limit_{Lo C}: Preset lower limit value for the center of the hit position
Limit_{Up C}: Preset upper limit value at the center of the hit position
[0128]
If this condition is satisfied, the image processing apparatus 11 determines that the center of the contact position of the piston ring 4 is good, and if this condition is not satisfied, the contact position of the piston ring 4 is determined. It is determined that the center of is defective.
[0129]
Thereafter, the image processing apparatus 11 determines the distance L between the lines as shown in FIG._CThe distance L having the maximum value_CAnd select this as the maximum value L_{C Max}And the distance L between the lines._CTo select the distance LC having the smallest value, which is the minimum value L_{C Min}After that, the calculation shown in the following formula is performed to determine the deflection width A at the center of the hit position._LAfter asking
A_L= L_{C Max}-L_{C Min}
It is calculated whether or not the condition shown in the following equation is satisfied.
AL <Limit_{Up CFl}
However, Limit_{Up CFl}: Predetermined upper limit value of the center of contact position
[0130]
If this condition is satisfied, the image processing apparatus 11 determines that the swing width A at the center of the contact position of the piston ring 4 is reached._LIf this condition is not satisfied, the deflection width A at the center of the contact position of the piston ring 4 is determined._LIs determined to be defective.
[0131]
Thereafter, the image processing device 11 generates a control signal indicating a non-defective product for the piston ring 4 determined to be good in all the above-described inspections, and the piston determined to be defective in any of the above-described inspections. A control signal for instructing the defective product to the ring 4 is generated and supplied to the imaging mechanism 2 to select the piston ring 4 (step ST23).
[0132]
<Display Image Creation Processing (Step ST24)>
When the inspection result display screen is designated by the personal computer device 13, the image processing device 11 uses the data obtained from the above-described inspection results, as shown in FIG. The piston image made easy to see by processing this image, the pass / fail judgment result obtained by inspecting the piston ring 4 to classify it as one of good, defective and semi-good, and the width of the contact surface on the piston ring 4 Presence / absence and position thereof, presence / absence of contact of the contact surface on the piston ring 4 and presence / absence of a defect (remaining width) between the contact surface and the chamfering position on the piston ring 4 and position thereof The presence / absence of smoothing processing on the left side of the contact surface in the piston ring 4 and its position, and the piston ring 4 Presence / absence of smoothing treatment on the right side of a certain contact surface and its position, the total number of piston rings 4 inspected so far, the total number of non-defective items in the piston ring 4 inspected so far, and in the piston ring 4 inspected so far Screen data for displaying the total number of semi-defective products and the total number of defective products in the piston ring 4 that have been inspected so far are generated and displayed on the respective monitor devices 8 and 12.
[0133]
Further, when an inspection status display screen is designated by the personal computer device 13, the image processing device 11 uses the data obtained from the above-described inspection results to use the piston ring 4 currently being inspected as shown in FIG. The piston image made easy to see by processing this image, the pass / fail judgment result obtained by inspecting the piston ring 4 to classify it as one of good, defective and semi-good, and the width of the contact surface on the piston ring 4 Presence / absence and position thereof, presence / absence of contact of the contact surface on the piston ring 4 and presence / absence of a defect (remaining width) between the contact surface and the chamfering position on the piston ring 4 and position thereof And the presence / absence of the smoothing process on the left side of the contact surface in the piston ring 4 and the position thereof, and the piston ring 4 The presence / absence and position of the smoothing process on the right side of the contact surface, the brightness of the image of the piston ring 4, the contrast of the image of the piston ring 4, the edge strength of the image of the piston ring 4, and the piston ring 4 Screen data for displaying the sharpness of the image is generated and displayed on each of the monitor devices 8 and 12 (step ST24).
[0134]
As described above, in this embodiment, the piston ring 4 is captured by the imaging mechanism and the outer peripheral portion is imaged while rotating the piston ring 4, and the image obtained by the imaging operation is processed by the control processing mechanism 3, The piston ring 4 is used as an ITV camera because it performs free gap detection, outer edge detection, chamfer position detection, contact surface detection, pass / fail determination, display image creation, etc. When the image of the outer periphery of the piston ring is captured by the line sensor 11, whether the piston ring 4 is good or not can be determined with high accuracy even if the piston ring image is distorted or the contrast of the piston ring image is small.
[0135]
【The invention's effect】
As described above, in the inventions according to claims 1 to 3, even when the image of the outer periphery of the annular member (piston ring) is taken from the camera, the annular member (piston ring) image is distorted. Even if the contrast of the annular member (piston ring) image is small, the quality of the annular member (piston ring) can be determined with high accuracy.
[0136]
In the invention according to claim 4, image processing of the annular member (piston ring) to be inspected can be facilitated, and an image of the annular member (piston ring) can be easily seen.
[0137]
In the invention according to claim 5, the processing accuracy of the image processing of the annular member (piston ring) to be inspected can be increased.
[0138]
Furthermore, in the invention described in claim 6, a portion to be detected in the image of the annular member (piston ring) to be inspected can be detected with high accuracy.
[Brief description of the drawings]
FIG. 1 is a block diagram showing an embodiment of an annular member outer periphery inspection device and an image processing method according to the present invention.
FIG. 2 is a side view showing an example of a piston ring inspected by the annular member outer periphery inspection device shown in FIG. 1;
FIG. 3 is a schematic diagram showing an example of inspection items of the annular member outer periphery inspection device shown in FIG. 1;
4 is a side view showing an example of a piston ring image capturing operation of the inspection station apparatus shown in FIG. 1. FIG.
5 is a top view showing an example of a piston ring image capturing operation of the inspection station apparatus shown in FIG. 1. FIG.
6 is a schematic diagram illustrating an example of content of an image signal output from the imaging mechanism illustrated in FIG. 1. FIG.
FIG. 7 is a flowchart showing an example of an inspection operation of the annular member outer periphery inspection apparatus shown in FIG. 1;
FIG. 8 is a flowchart showing an example of an inspection operation of the annular member outer periphery inspection apparatus shown in FIG. 1;
FIG. 9 is a flowchart showing an example of an inspection operation of the annular member outer periphery inspection apparatus shown in FIG. 1;
10 is a schematic diagram showing an example of an effective image cutting operation performed in the free gap detection process of the control processing mechanism shown in FIG.
11 is a schematic diagram showing an example of a horizontal projection processing area cut-out operation performed in the free gap detection process of the control processing mechanism shown in FIG. 1; FIG.
12 is a schematic diagram showing an example of horizontal projection calculation operation performed in the free gap detection process of the control processing mechanism shown in FIG.
FIG. 13 is a schematic diagram showing an example of threshold calculation for the outer edge search performed in the free gap detection process of the control processing mechanism shown in FIG. 1;
14 is a schematic diagram showing an example of an outer edge search operation performed in the free gap detection process of the control processing mechanism shown in FIG. 1. FIG.
15 is a schematic diagram showing an example of a free gap detected by a free gap detection process of the control processing mechanism shown in FIG. 1. FIG.
16 is a schematic diagram showing an example of a free gap detected by a free gap detection process of the control processing mechanism shown in FIG.
17 is a schematic diagram showing an example of an outer edge detected by an outer edge detection process of the control processing mechanism shown in FIG. 1;
18 is a schematic diagram showing an example of a threshold value calculation operation performed in the outer edge detection process of the control processing mechanism shown in FIG. 1. FIG.
FIG. 19 is a schematic diagram showing an example of a left outer edge search operation performed in an outer edge detection process of the control processing mechanism shown in FIG. 1;
20 is a schematic diagram showing an example of determination of the left outer edge performed in the outer edge detection process of the control processing mechanism shown in FIG. 1. FIG.
FIG. 21 is a schematic diagram showing an example of a right outer edge search operation performed in an outer edge detection process of the control processing mechanism shown in FIG. 1;
22 is a schematic diagram showing an example of a processing region determined in an outer edge detection process of the control processing mechanism shown in FIG. 1. FIG.
FIG. 23 is a schematic diagram showing an example of division of the processing target area performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1;
24 is a schematic diagram showing an example of a vertical projection data calculation operation performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1; FIG.
FIG. 25 is a schematic diagram showing an example of a search operation for a chamfer position and a peak position performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1;
26 is a schematic diagram showing an example of a chamfer position and a peak obtained by the chamfer position detection process of the control processing mechanism shown in FIG. 1. FIG.
FIG. 27 is a schematic diagram showing an example of a moving average process performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1;
FIG. 28 is a schematic diagram showing an example of a minimum position detection operation performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1;
29 is a schematic diagram showing an example of a temporary search range of the chamfer position obtained by the chamfer position detection process of the control processing mechanism shown in FIG. 1. FIG.
30 is a schematic diagram showing a determination example of a chamfer position search range performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1; FIG.
FIG. 31 is a schematic diagram showing an example of determining a chamfer position search range performed in the chamfer position detection process of the control processing mechanism shown in FIG. 1;
32 is a schematic diagram showing an example of a chamfer position correcting operation obtained by the chamfer position detection process of the control processing mechanism shown in FIG. 1; FIG.
33 is a schematic diagram showing an example of a chamfer position performed in a chamfer position detection process of the control processing mechanism shown in FIG. 1. FIG.
34 is a schematic diagram showing an example of a hit position search range performed in the hit face detection process of the control processing mechanism shown in FIG. 1; FIG.
FIG. 35 is a schematic diagram showing an example of division of a search range of a hit position performed in a hit face detection process of the control processing mechanism shown in FIG. 1;
36 is a schematic diagram showing an example of connection between small sections performed in the contact surface detection process of the control processing mechanism shown in FIG. 1;
FIG. 37 is a schematic diagram showing an example of an operation for calculating an evaluation value used in a connecting operation between small sections performed in a contact surface detection process of the control processing mechanism shown in FIG. 1;
38 is a diagram showing an example of a general table used in a connecting operation between small sections performed in the contact surface detection process of the control processing mechanism shown in FIG. 1;
FIG. 39 is a schematic diagram showing an example of a weight function creation operation performed in the contact surface detection process of the control processing mechanism shown in FIG. 1;
40 is a schematic diagram showing an example of a weight table created by the contact surface detection process of the control processing mechanism shown in FIG. 1. FIG.
41 is a schematic diagram showing an example of a search range of a hit position with respect to the top ring performed in the hit face detection process of the control processing mechanism shown in FIG. 1;
42 is a schematic diagram showing an example of a search range of a hit position with respect to the second ring performed in the hit face detection process of the control processing mechanism shown in FIG. 1; FIG.
43 is a schematic diagram showing an example of search range dividing operation performed in the contact surface detection process of the control processing mechanism shown in FIG. 1; FIG.
44 is a schematic diagram showing an example of a contact position correcting operation performed in the contact surface detection process of the control processing mechanism shown in FIG. 1; FIG.
45 is a schematic diagram showing an example of a contact position correcting operation performed in the contact surface detection process of the control processing mechanism shown in FIG. 1; FIG.
46 is a schematic diagram showing an example of a contact position correcting operation performed in the contact surface detection process of the control processing mechanism shown in FIG. 1; FIG.
47 is a schematic diagram showing an example of an operation for determining a smoothing location at a contact position performed in the contact surface detection process of the control processing mechanism shown in FIG. 1;
48 is a schematic diagram illustrating an example of an operation for determining a smoothing location at a contact position performed in the contact surface detection process of the control processing mechanism illustrated in FIG. 1;
49 is a schematic diagram showing an example of a smoothing target section of a hit position performed in the hit face detection process of the control processing mechanism shown in FIG. 1. FIG.
50 is a schematic diagram illustrating an example of a boundary portion of a smoothing operation at a contact position performed in a contact surface detection process of the control processing mechanism illustrated in FIG. 1. FIG.
FIG. 51 is a schematic diagram showing an example of a smoothing operation of a hit position performed in a hit face detection process of the control processing mechanism shown in FIG. 1;
52 is a perspective view showing an example of a piston ring processed in the determination process of the control processing mechanism shown in FIG. 1. FIG.
53 is a schematic diagram showing an example of a remaining width determining operation example for the top ring performed in the determination process of the control processing mechanism shown in FIG. 1; FIG.
54 is a schematic diagram showing an example of the remaining width determination operation example for the second ring performed in the determination process of the control processing mechanism shown in FIG. 1; FIG.
55 is a schematic diagram showing an example of a center detection operation of a hit position performed in the determination process of the control processing mechanism shown in FIG. 1. FIG.
FIG. 56 is a schematic diagram showing an example of the operation for determining the center of a hit position performed in the determination process of the control processing mechanism shown in FIG. 1;
FIG. 57 is a schematic diagram showing an example of a determination operation of a contact position deflection width performed in the determination processing of the control processing mechanism shown in FIG. 1;
58 is a schematic diagram showing an example of a display screen displayed in the display image creation process of the control processing mechanism shown in FIG.
59 is a schematic diagram showing an example of a display screen displayed in the display image creation process of the control processing mechanism shown in FIG.
[Explanation of symbols]
1 Perimeter inspection system for annular members
2 Imaging mechanism
3 Control processing mechanism
4 Piston ring (annular member)
5 Transport control device
6 Inspection station equipment
7 Transport device
8 Image monitor device
9 Lighting power supply
10 Light guide
15 Holding mechanism
11 Line sensor (camera)
11 Image processing device
12 Image monitor device
13 Personal computer device
14 Serial printer device

Claims (6)

The outer peripheral portion of the annular member is imaged by rotating the annular member side to be inspected or revolving the camera side disposed at a predetermined distance from the annular member to generate an image signal of the outer peripheral portion of the annular member. An image pickup mechanism, and a control processing mechanism that inspects the outer periphery of the annular member based on the image signal of the annular member obtained by the image pickup mechanism,
The control processing mechanism includes:
Processing target area determining means for determining a processing target area based on an image signal of the annular member;
An image specifying means for detecting the outer edge of the annular member for each line in the determined processing target area and specifying the annular member image;
A chamfer position detecting means for detecting the chamfer position of the annular member by searching in the identified annular member image;
A contact surface detecting means for detecting a contact surface by searching a range corresponding to the detected chamfering position;
Determination means for determining the quality of the annular member based on the state of the detected contact surface;
An apparatus for inspecting the outer periphery of an annular member, comprising The annular member outer periphery inspection device according to claim 1, wherein the annular member is a piston ring, and the processing target region determination unit determines a processing target region by detecting a free gap of the piston ring. Rotate the annular member side to be inspected or revolve the camera side arranged away from the annular member by a predetermined distance to capture the outer peripheral part of the annular member to generate an image signal,
Based on the image signal of the annular member, determine the processing target region,
For each line in the processing target area, the outer edge of the annular member is detected to identify the annular member image,
Search the specified annular member image to detect the chamfer position of the annular member,
Search the range corresponding to the chamfering position of the annular member image to detect the contact surface,
Determining the quality of the annular member based on the state of the contact surface;
A method for inspecting the outer periphery of an annular member, characterized by: Capture the image signal of the annular member obtained by the imaging mechanism,
For each line orthogonal to the longitudinal direction of the annular member in the captured image signal, the left edge position and the right edge position of the annular member in the image signal are detected to recognize the contour of the annular member. ,
Detecting the center of the left edge position and the right edge position for each line;
After the detected center is set as the center position for each line, the image of the annular member is corrected to a straight shape by shifting the image signal for each line so that the center position of each line is linear. , Perform image processing on this image,
An image processing method of an annular member characterized by the above. In an image processing method for capturing an image signal of an annular member obtained by an imaging mechanism and processing the image signal,
When processing an image of an annular member based on the captured image signal, an initial threshold value obtained by adding a preset offset value to the pixel density value at the start position of each line is used as a reference. Toward the end position side, search for the density value of each pixel to be processed,
When the change rate of the density value of the pixel to be processed at each position is smaller than a preset value, the image signal is processed while updating the threshold value based on the change of the density value,
When the change rate of the density value of the pixel to be processed at each position is larger than a preset value, the updating of the threshold value up to that point is stopped and the image signal is processed.
An image processing method of an annular member characterized by the above. In an image processing method for capturing an image signal of an annular member obtained by an imaging mechanism and processing the image signal,
After obtaining the main density level based on the maximum density value and the minimum density value of each processing target pixel in the processing target area in the image signal, the maximum density value is set on one of two orthogonal axes. Plotting the main density level and the minimum density value, setting the main density level to be “1”, setting the maximum density value to a preset first weight function coefficient, Create a weight function table using a curve set as the second weight function coefficient for which the minimum density value is set in advance as a weight function,
When capturing the image signal and performing processing on the image of the annular member, the density value of the pixel to be processed that becomes a differential peak of the image signal is converted into a weight value using the weight function table, and Creating an evaluation function based on detecting a desired portion of the annular member;
An image processing method of an annular member characterized by the above.

Images