JP6207008B2 - Inspection apparatus and inspection method - Google Patents

Description

The present invention relates to an inspection technique, and more particularly to a technique for cutting out a partial image including an inspection range from an image obtained by photographing an object to be conveyed and performing an inspection based on the partial image.
Regarding technology.

  Conventionally, in a production factory line, the appearance of a product is inspected while the product is conveyed by a conveying device. This inspection has been carried out visually by an inspector, but there are problems such as a burden on the inspector and a variation in inspection accuracy due to the subject's inspection relying on subjectivity. Therefore, automatic inspection by image processing or the like has been proposed.

  For example, there are Patent Documents 1 to 4 as inspection techniques for PET bottles, and Patent Document 5 as an inspection technique for labels attached to containers. In any of these techniques, an object to be inspected is photographed with a camera to acquire an image, and a wrinkle or the like is inspected based on the image.

  In Patent Documents 1 and 2, the presence or absence of wrinkles is inspected by so-called template matching, in which an image of a test part without wrinkles is compared with an image of a test part of an inspection object. On the other hand, in Patent Documents 3 to 5, the image of the inspection region of the inspection object is binarized, and the presence or absence of wrinkles is inspected based on the area.

JP-A-2005-308441 JP 2007-113933 A JP 2007-57431 A JP 2008-157858 A JP 2008-281477 A

  With the technique of the patent document mentioned above, it can test | inspect automatically the presence or absence of the flaw of a to-be-inspected object by image processing. However, in these techniques, it is assumed that the examination site is in the same position in the image. For example, many products are flowing on the production line, and the distance and angle from the camera differ depending on which position on the conveyor belt the product flows. In such a case, the inspection site in the image is displaced from the assumed position, and the inspection accuracy may be reduced. In particular, in the template matching, the inspection accuracy depends on the accuracy of the position where the template is arranged. In order to eliminate the dependency of the position accuracy, it is necessary to perform template matching while changing the template position in the image, but there is a problem that the processing time increases.

  In view of the above problems, an object of the present invention is to provide an inspection technique capable of accurately inspecting even if the position of an inspection site in an image changes.

In order to solve the above-described problem, an inspection apparatus according to the present invention sets an imaging unit that captures an image of an object to be transported and acquires an image, and an inspection region that includes an inspection site of the object to be inspected in the image. An inspection area setting unit, a binary image generation unit that binarizes pixel values of the image included in the inspection area and generates a binary image, and an area having a predetermined one pixel value in the binary image An inspection mask generation unit that generates an inspection mask based on the image, and sets the inspection mask for the image, and sets the inspection mask in accordance with the pixel value of the pixel included in the inspection mask in the image. A heel determination unit that determines presence or absence, and the imaging unit is installed at an elevation angle position having an elevation angle or an depression position having a depression angle with respect to the examination site protruding laterally with respect to the inspection object. The inspection mask generator generates the imaging. Part upward by contracting the lower end of the test mask when in the elevation position, when the imaging unit is in the depression angle position to contract the upper end of the inspection mask downward.

  In this configuration, an inspection area is set in an image obtained by imaging the inspection object, and the presence or absence of wrinkles is inspected based on the pixels in the inspection area. Can do. In addition, since the inspection mask is generated based on the region in the binary image obtained by binarizing the pixels in the inspection region and the presence or absence of wrinkles is determined based on the pixels in the inspection mask, the inspection region is included in the inspection region. Even if a part or background other than is included, it is possible to accurately determine wrinkles of the examination part.

  In this configuration, the inspection mask is contracted according to the position of the imaging unit with respect to the inspection region, so that the lower surface or the upper surface of the inspection region is not included. Therefore, if the presence or absence of wrinkles is determined based on the pixel values of the pixels included in such an inspection mask, wrinkles on the side surface of the inspection site can be determined.

The technical features of the inspection apparatus of the present invention can also be applied to similar inspection methods. For example, an inspection area setting step for setting an inspection area including an inspection site of the inspection object in an image of the inspection object to be conveyed imaged by the imaging unit, and a pixel value of the image included in the inspection area A binarization step for generating a binary image, an inspection mask generation step for generating an inspection mask based on a region having a predetermined one pixel value in the binary image, and the image set the test mask Te, and a flaw determining step determines the presence or absence of defects in the inspection object based on the pixel values of pixels included in the inspection mask in the image, the imaging unit is the object It is installed at an elevation angle position having an elevation angle or an depression angle position having a depression angle with respect to the examination site protruding laterally with respect to the inspection object, and when the imaging unit is at the elevation angle position, the lower end of the inspection mask is Upward By condensation, when the imaging unit is in the depression angle position can be used in the inspection method of contracting the upper end of the inspection mask downward.

It is an external view of the inspection apparatus of the present invention. It is a functional block diagram of the inspection apparatus of the present invention. It is a flowchart showing the flow of a process of the inspection apparatus of this invention. It is an example of the image which image | photographed the test | inspection site | part of the to-be-inspected object. It is the figure which set the test | inspection area | region to the image of FIG. 6 is a binary image obtained by performing binarization processing on the inspection region in FIG. 5. 6 is a flowchart illustrating a flow of processing for generating an inspection mask in the first embodiment. It is an image showing the inspection mask obtained from the binary image of FIG. It is an image showing the inspection mask which correct | amended the inspection mask of FIG. It is a flowchart showing the flow of a wrinkle determination process. This image clearly shows the position of the heel. 6 is a flowchart illustrating a flow of inspection mask generation processing in the second embodiment. 6 is an image showing a corrected inspection mask in Example 2.

  Embodiments of the inspection apparatus of the present invention will be described below with reference to the drawings. In the present embodiment, the inspection object will be described as a plastic bottle, the inspection site as a neck ring, and the bag as a neck ring bag. Of course, the inspection object and the inspection part are not limited to these, and the inspection object can be a container such as a bottle or a packaging material such as cardboard, and the inspection part is attached to the cap or the inspection object. Labels and stickers can also be used. Further, wrinkles indicate an abnormality that has occurred in an examination site. When the examination site is a label or a seal, it includes not only wrinkles but also abnormal conditions such as peeling.

〔overall structure〕
FIG. 1 is an external view of an inspection apparatus according to the present invention, and FIG. 2 is a functional block diagram of the inspection apparatus. As shown in FIG. 1, a plastic bottle P, which is an object to be inspected, is placed on the belt of the transport device C and transported in the right direction in the figure. A sensor S for detecting the plastic bottle P is provided on the transport path of the transport apparatus C. The output of the sensor S is input to the control unit 1 described later.

  In addition, a camera 2 that captures a neck ring of a plastic bottle P that is an inspection site is installed on the side of the transport device C at the same position as the sensor S in the transport direction. In the present embodiment, the camera 2 is installed below the neck ring so that the optical axis faces the neck ring. Therefore, the camera 2 in this embodiment has an elevation angle with respect to the neck ring. The camera 2 is connected to the processing device T.

[Processing equipment]
The processing device T comprises a general-purpose computer, and each functional unit is configured by software. Of course, each functional unit may be configured by hardware or software and hardware.

  As illustrated in FIG. 2, the processing device T includes a control unit 1 that performs various controls, an image acquisition unit 3 that acquires an image based on the output of the camera 2, and an inspection that sets an inspection region in the image acquired by the image acquisition unit 3. An area setting unit 4, a binary image generation unit 5 that generates a binary image obtained by binarizing pixel values of pixels included in the inspection area, and an inspection based on an area having a predetermined one pixel value in the binary image An inspection mask generation unit 6 that generates a mask and a wrinkle determination unit 7 that determines the presence or absence of wrinkles based on the pixel values of pixels included in the inspection mask are provided.

  The control unit 1 has a function of performing various controls. For example, when receiving the detection signal of the plastic bottle P by the sensor S, the camera 2 and the image acquisition unit 3 are instructed to acquire an image. Further, when the bottle determining unit 7 determines that there is a bottle, the operator discharges the plastic bottle P determined to have a bottle to a discharge unit (not shown) that notifies the operator to that effect. Control such as issuing instructions.

  As described above, the camera 2 photographs the neck ring of the plastic bottle P from below with an elevation angle. In this embodiment, the camera 2 uses a monochrome digital still camera, but a color digital still camera or digital movie camera may be used. An analog camera may be used. In this case, the image acquisition unit 3 needs an A / D conversion function.

  The image acquisition unit 3 has a function of receiving an output from the camera 2 and acquiring an image. As described above, since the camera 2 is a monochrome digital still camera in this embodiment, the image acquisition unit 3 of this embodiment simply functions as an input unit for image data output from the camera 2. If the camera 2 is a color digital still camera, the image acquisition unit 3 converts the color image data acquired from the camera 2 into a grayscale image. If the camera 2 is a digital movie camera, the image acquisition unit 3 extracts a predetermined frame from the moving image data. The camera 2 and the image acquisition unit 3 constitute a photographing unit of the present invention.

  The inspection area setting unit 4 has a function of setting an inspection area in the image acquired by the image acquisition unit 3. The inspection region is a region including an inspection part in the image and is a region to be processed later. The inspection area can be set by the operator while viewing the captured image, but in this embodiment, an inspection area having a preset position and size is set. As described above, since the plastic bottle P is placed and transported on the belt of the transport device C, the positional deviation is suppressed to a predetermined value or less. Therefore, the position and size of the inspection region may be set so as to include the neck ring position in the image taken in a state where the PET bottle P is shifted to both ends in the direction orthogonal to the conveyance direction.

  The binary image generation unit 5 has a function of binarizing the pixel values of the pixels included in the examination area among the pixels of the image acquired by the image acquisition unit 2. Binarization sets the pixel value of a pixel to one pixel value (for example, 0) if the pixel value of the pixel is less than or equal to a predetermined threshold, and if the pixel value of the pixel is greater than the predetermined threshold, In this process, the pixel value of the pixel is set to the other pixel value (for example, 1). Therefore, if the threshold value can be determined, the image can be binarized. Various methods can be used for determining the threshold. For example, it may be obtained experimentally or by a statistical method such as discriminant analysis.

  The inspection mask generation unit 6 has a function of generating an inspection mask from the binary image generated by the binary image generation unit 5. The inspection mask is a mask image used for masking pixels other than the pixels corresponding to the inspection region when determining wrinkles. The inspection mask generation unit 6 divides the binary image into regions each having the same pixel value and is composed of adjacent pixels, and generates an inspection mask from the region division result.

  The wrinkle determination unit 7 has a function of setting an inspection mask for an image and determining the presence or absence of wrinkles at the inspection site based on the pixel values of the pixels included in the inspection mask in the image.

  FIG. 3 is a flowchart showing the flow of processing of the inspection apparatus in the present embodiment. When the sensor S detects the plastic bottle P transported on the transport apparatus C (Yes branch of # 01), a signal from the sensor S is input to the control unit 1. In response to the signal, the control unit 1 transmits an instruction to the camera 2 and the image acquisition unit 3 to acquire an image, and the camera 2 and the image acquisition unit 3 receive the instruction and perform necking of the plastic bottle P. The image containing is acquired (# 02). FIG. 4 is an example of an image photographed in this way. As described above, in the present embodiment, since the camera 2 is installed so as to have an elevation angle below the neck ring, the lower surface and side surfaces of the neck ring are included in the image. In addition, the neck ring has a ridge D at a position surrounded by a circle in the drawing. Of course, the installation position of the camera 2 can be changed as appropriate according to the inspection object and the inspection site.

  The inspection area setting unit 4 sets the inspection area R for the image photographed as described above (# 03, see FIG. 5). As is apparent from the figure, the height of the inspection region R is set to be larger than the height of the neck ring in the image. Accordingly, the neck ring can be included in the inspection region R even if the position of the neck ring in the image is shifted due to the position shift of the plastic bottle P on the transport device C.

  The binary image generation unit 5 performs a binarization process on the pixels included in the inspection region R in the image to generate a binary image (# 04, see FIG. 6). In this embodiment, this binary image is obtained as an image having the same size as the inspection region R.

  As described above, the binarization threshold can be obtained by various methods. Statistical methods such as discriminant analysis can accurately determine the threshold even if the desired threshold fluctuates, but there is a problem that the amount of calculation increases. On the other hand, when the desired threshold value varies with the fixed threshold value, the inspection accuracy may be reduced. Therefore, it is preferable to determine the threshold value by applying the discriminant analysis method within a predetermined range centered on the fixed threshold value.

  As shown in FIG. 5, the neck ring has a lower luminance value than the other parts. For this reason, in this embodiment, as shown in FIG. 6, binarization is performed by setting a pixel value equal to or less than the threshold to 1 (white) and a pixel value greater than the threshold to 0 (black).

  The inspection mask generation unit 6 generates an inspection mask based on the binary image generated by the binary image generation unit 5 (# 05). Specifically, the process of the flowchart of FIG. 7 is performed.

  First, the inspection mask generation unit 6 performs a labeling process on a pixel having one pixel value in the binary image (# 11). As described above, in the present embodiment, the neck ring that is the inspection site is expressed as a white pixel in the binary image. Therefore, here, labeling is performed on white pixels in the binary image. Of course, when the examination region is expressed as a black pixel in the binary image, the black pixel in the binary image is labeled. For example, the binary image in FIG. 6 includes two areas m1 and m2, and these areas can be identified by labeling.

  Here, comparing FIG. 5 with FIG. 6, it can be seen that the region m1 is a neck ring and the region m2 is a cap. Therefore, in order to increase the accuracy of the subsequent wrinkle determination process, it is desirable to include only the region m1 corresponding to the neck ring in the inspection mask. Therefore, the inspection mask generation unit 6 excludes a region that does not correspond to the inspection region (region m2 in this example) from the inspection mask (# 12 to # 15).

  The inspection mask generator 6 calculates the feature amount of each area obtained by labeling (# 12). As the feature amount, various values can be used as long as they are values that can represent an examination site, such as an area, a fillet diameter, and an aspect ratio of a circumscribed rectangle.

  The inspection mask generation unit 6 selects one region and compares the feature amount of the region with the feature amount of a binary image representing a preset inspection region, so that the region is suitable as an inspection mask. (# 13). If it is determined that the image is not suitable, the area is deleted from the binary image (the pixel value is set to 0) (# 14). In the case of the above-described example, the area m2 is excluded from the binary image of FIG. 6, and the inspection mask shown in FIG. 8 is obtained. This process is performed for all areas (# 12 to # 15).

  Further, the inspection mask generation unit 6 corrects the inspection mask obtained by the above-described process (# 16). Although this correction processing is not essential, the inspection mask obtained by the processing so far is simply generated on the basis of the binary image. Therefore, due to causes such as noise at the time of photographing and deviation of the binarization threshold, There is a possibility that a hole is included in the mask or the outline of the inspection mask is shifted from the inspection site. In order to compensate for this, the contour line of the inspection mask is corrected in this embodiment.

  In this embodiment, the inspection mask is expanded and contracted. In this way, the inspection mask shown in FIG. 9 is obtained from the inspection mask of FIG. As is clear from comparison of these figures, the contour line above the inspection mask becomes a smooth curve by correction. Of course, the inspection mask correction method is not limited to this, and other methods such as averaging + binarization may be used.

  When the inspection mask is generated in this way, the eyelid determining unit 7 sets the inspection mask on the image, and based on the pixel value of the pixel in the image corresponding to the coordinates of the pixel whose pixel value of the inspection mask is 1. To determine whether or not (# 06).

  As described above, in this embodiment, the binary image is an image having the same size as that of the inspection region R, and thus the image representing the inspection mask is also the same size as the inspection region. Therefore, when setting the inspection mask on the image, the inspection mask and the inspection area are set to coincide with each other.

In the present embodiment, the presence or absence of wrinkles is determined by the following processing, and a wrinkle image representing wrinkles is generated. First, an initialization process is performed prior to the process (# 21). In this embodiment, the inspection site pixel number N indicating the number of pixels indicating the inspection region and the 疵 pixel number N _D indicating the number of pixels indicating 疵 are initialized to 0 and the 疵 image indicating 疵 is initialized (all The pixel value is set to 0).

  Subsequently, as preprocessing, the pixel values of all the pixels corresponding to the coordinates of the pixel whose inspection mask pixel value is 1 among the pixels in the image are acquired, and the statistics are calculated (# 22). In this embodiment, the average and standard deviation are used as statistics.

  Next, among the pixels in the image, one pixel corresponding to the coordinates of the pixel whose pixel value of the inspection mask is 1 is selected (# 23), and an evaluation value based on the statistic is calculated for that pixel (# 23). # 24). In this embodiment, a deviation value is used as the evaluation value. At this time, the inspection site pixel number N is also incremented.

The evaluation value of the selected pixel is compared with a predetermined condition, and it is determined whether or not the pixel is habit (# 25). In this embodiment, the standard deviation is determined as a flaw if the lower limit threshold value or less or the upper threshold or more, increments the flaw pixel number N _D (# 26). In addition, the pixel value of the pixel determined to be 疵 is set as the pixel value of the corresponding pixel of the 疵 image (# 27).

When this process is performed for all the pixels (# 23 to # 28), the wrinkle determination unit 7 determines whether or not there is a wrinkle in the examination site (# 29). In this embodiment, if the ratio of the number of defective pixels N _D with respect to the inspection site pixel number N is equal to or greater than a predetermined threshold value, it is determined that there is a flaw.

  Of course, the method for determining wrinkles is not limited to this, and other methods may be used.

  The determination result is transmitted to the control unit 1. When the control unit 1 acquires information that wrinkles have occurred in the examination site, the control unit 1 may control the plastic bottle P to be excluded from the transport device, or notify the operator that there is wrinkles. You may do it. Further, when notifying the operator, it is preferable to present an image clearly showing the wrinkle position as shown in FIG. 11 because it helps the operator to check the wrinkle visually. This image can be created by using the haze image created by the haze determination process.

  The processing in this embodiment is different from the processing in the first embodiment only in the inspection mask correction processing. Therefore, only the inspection mask correction process will be described.

  As described above, in this embodiment, the camera 2 is installed so as to have an elevation angle from below the neck ring. Therefore, the photographed image includes the bottom surface and side surface of the neck ring. If the wrinkle is determined based on such an image, there is a risk of overlooking the wrinkle generated only on the side surface of the neck ring. In particular, when determining wrinkles by area ratio as in the first embodiment, wrinkles generated only on the side surfaces of the neck ring are easily overlooked. For this reason, in this embodiment, the inspection mask is corrected so as to conform to the pixel indicating the side surface of the neck ring.

  FIG. 12 is a flowchart showing the flow of inspection mask generation processing in this embodiment. # 31 to # 36 are the same as # 11 to # 16 in FIG. In this embodiment, the inspection mask (see FIG. 9) obtained in this way is corrected again.

  In this embodiment, the lower part of the inspection mask corresponds to the lower surface of the neck ring, and the upper part corresponds to the side surface of the neck ring. Therefore, the lower end of the inspection mask is contracted upward by a predetermined percentage (# 37). Thereby, an inspection mask corresponding to the side surface of the neck ring is obtained (see FIG. 13). The rate of contraction is appropriately set according to the size of the neck ring and the elevation angle of the camera 2. Further, the rate of contraction need not be constant. For example, the vicinity of the center may be set small and the vicinity of the end may be set large.

  In this embodiment, since the camera 2 is installed at a position having an elevation angle, the lower end of the inspection mask is contracted upward. However, when the camera 2 is installed at a position having a depression angle, the upper end of the inspection mask. Is contracted downward.

  INDUSTRIAL APPLICABILITY The present invention can be used for inspection technology for containers such as PET bottles to be conveyed, bags for packaging materials such as corrugated cardboard, and labels and seals attached to these.

P: PET bottle (inspection object)
R: Inspection area 2: Camera (imaging part)
3: Image acquisition unit (shooting unit)
4: Inspection region setting unit 5: Binary image generation unit 6: Inspection mask generation unit 7: Wrinkle determination unit

Claims (2)

A photographing unit for photographing an object to be conveyed and acquiring an image;
An inspection region setting unit for setting an inspection region including an inspection part of the inspection object in the image;
A binary image generation unit that binarizes pixel values of the image included in the inspection area and generates a binary image;
An inspection mask generation unit that generates an inspection mask based on a region having a predetermined one pixel value in the binary image;
A wrinkle determination unit that sets the inspection mask for the image and determines whether or not the inspection object has wrinkles based on a pixel value of a pixel included in the inspection mask in the image ,
The imaging unit is installed at an elevation angle position having an elevation angle or an depression angle position having a depression angle with respect to the examination site protruding laterally with respect to the inspection object,
The inspection mask generation unit contracts the lower end of the inspection mask upward when the imaging unit is in the elevation position, and contracts the upper end of the inspection mask downward when the imaging unit is in the depression position. Inspection apparatus characterized by that. An inspection region setting step for setting an inspection region including an inspection region of the inspection object in an image of the inspection object to be conveyed imaged by the imaging unit;
A binarization step of binarizing pixel values of the image included in the inspection area to generate a binary image;
An inspection mask generating step for generating an inspection mask based on an area having one predetermined pixel value in the binary image;
A wrinkle determination step that sets the inspection mask for the image, and determines whether or not the inspection object has wrinkles based on a pixel value of a pixel included in the inspection mask in the image ,
The imaging unit is installed at an elevation angle position having an elevation angle or an depression angle position having a depression angle with respect to the examination site protruding laterally with respect to the inspection object,
An inspection method, wherein the lower end of the inspection mask is contracted upward when the imaging unit is at the elevation position, and the upper end of the inspection mask is contracted downward when the imaging unit is at the depression position. .

Images