JP3098984B2 - Inspection system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a board inspection system, and more particularly to a system for inspecting a building board running at a high speed on a building board production line without stopping the line.

[0002]

2. Description of the Related Art Conventionally, under illumination conditions using a standard light source or so-called north window light with reduced brightness, the quality of a painted building board is judged visually by comparing it with a comparative reference product. In this case, the judgment items are mainly judgments of hue defects, density (depending on the amount of paint applied) defects and dirt, and in the case of sputter coating, the size of spots and the scattered state are also determined. .

[0003] There have been developed systems in which hue and density are checked using a colorimetric system using a precision spectrophotometer, and an image is enlarged using a microscope and projected on a television screen. They are used to measure a test plate cut to a predetermined size in a stationary state, and stop the line immediately after finding a defective coating plate during processing, or selectively remove only defective plates. It has not been applied to the field of on-board inspection boards. That is, the measurement range of the optical equipment to be used is limited to the spot range, the measurement distance is short, etc., it has a large area, and it can not cope with measuring the entire surface of the building board that is running It is thought that this is the cause. Further, it is inevitable that a system for moving the optical device so as to roughly scan the building board can be considered. However, since vibration due to movement is a great enemy of stable measurement for the optical device, the realization of such a system is extremely difficult. You are in a difficult situation.

On the other hand, there is a simple line monitoring system to which an imaging system using a television camera is applied, and there are many examples of the system. However, the system is intended only to monitor the occurrence of abnormalities in the line, and when applied to the inspection board surface, it has never been satisfactory in terms of reading accuracy. Absent. That is, the traveling speed of the building board is considerably high (40 to
The reason is that the reading ability cannot be adapted to the case of 60 m / min). Furthermore, even for a test board which targets a fine spot pattern such as a sputter coating, it cannot be expected that the reading and imaging accuracy of such a degree that the size and the scattered state of the spot can be determined.

[0005]

SUMMARY OF THE INVENTION The present invention basically uses a normal television camera image pickup system and processes image data acquired online using a personal computer to obtain a high-accuracy building board inspection. It is an object of the present invention to provide a plate inspection system that can perform plate inspection.

[0006]

SUMMARY OF THE INVENTION A board inspection system according to the present invention is for inspecting a building board. A TV camera for imaging the running building board and an image signal from the TV camera are provided.
Of which the same part of the building board was imaged with different frames
And image processing means for processing an image signal by removing any of the signals .

Further, it is preferable that the image processing means performs predetermined image signal processing during a time when the removed signal is present, since the image signal processing can be performed at high speed. In addition, since the image processing means inspects the contour shape of the building board, it is possible to inspect whether or not there is a chip at an end of the building board.

[0008] The image processing means may include a
By detecting spots, the painting status of the building board
Can be inspected. Further, the image processing means includes:
The number of the spots, the size of the spots or the scattering of the spots
By evaluating the situation, the building board
It can inspect the painting situation of

Further, the inspection board system of the present invention is for inspecting a building board, and a TV for imaging the running building board.
Camera and groove data from image signals from the TV camera
Image processing means for processing the image signal from which the color has been removed to inspect the hue of the building board . In addition, the T
That the V camera can change the imaging magnification,
This is preferable because a more detailed inspection can be performed.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the drawings. In the present invention,
As shown in FIG. 1, an image of a building board 2 is captured by a TV camera 1. The TV camera 1 sets the imaging distance so that the width m in the longitudinal direction of the imaging screen 3 slightly overlaps the end guides 4 that guide the building board 2. Fine adjustment of both end images is performed at the stage of image processing. The start and end of the sampling of the imaging data of the building board 2 are defined by detecting the front and rear ends of the building board 2 with the photoelectric switch 9 (shown in FIG. 6).

FIG. 2 is a diagram (part 1) for explaining the normal image processing in the inspection board system according to the present invention. Building board 2
And the transmission relationship of the image frames captured by the TV camera 1. Assuming that the traveling speed v of the building board 2 is v = L [m / sec], 30 image frames are captured for one second, that is, while the building board 2 moves the distance L. However, in the second and subsequent image frames to be transmitted, the imaging areas of the building board 2 overlap.

FIG. 3 is an explanatory view (No. 2) of the normal video processing. A method of obtaining a scanning line number corresponding to the rear end of the overlapping portion of the imaging area of the building board 2 in the second and subsequent image frames will be described. The scanning line number is required to specify the sampling time of the image data. The start time will be aligned by specifying the input at the gate. As described above, since the traveling speed v of the building board 2 is v = L [m / sec], the captured image P ₁ + d ₁ + d ₂ + of the building board L × m is obtained.
· · · + D _n rectangular regions d ₁ of, d _2, ···, each of the first line number x of d _n is, x: 480 = (l- L / 30): l x = 480 (l-L / 30) / l. That is, the rectangular area d _1, d _2, · · ·, for capturing images of d _n is, by performing the sampling of the image data to the frame area up to 480 onward from the line corresponding to the x value determined here good.

FIG. 4 is an explanatory diagram (part 3) of the normal video processing. 4 shows a relationship between a transmission frame and sampling of image data. By combining the sampled image data, image data for all building boards 2 is obtained. Also, sampling will have an interruption period,
By performing image processing using this time, by the time sampling of one building board 2 ends, the image processing can be almost completed. Therefore, as a result, image processing can be performed at high speed.

In this embodiment, the first captured image is set to the full screen, and the subsequent captured images are defined as thin rectangular areas d ₁ , d ₂ ,.
.., D _n , but if the length of the building board 2 is determined, sampling may be performed so that all captured images including the first captured image have the same width.

FIG. 5 is a diagram for explaining sampling and encoding of image data. Here, the sampling frequency is 13.5 MHz and the number of effective sampling points is 480 * 640 as a model for simplicity, but it is of course possible to increase the resolution. In addition, the images of the guides 4 on both ends of the building board 2 are excluded as unnecessary data in the data processing process. Therefore, in the description here, 480 * 64 is set as the number of pixels forming the image of the building board 2 portion.
Use 0. The obtained RGB data can be quantized as it is using a CCD camera.

As shown in FIG. 5 (a), the sampling structure is a square lattice sampling structure, and the encoding format digitally processes the luminance signal Y and the color signals RY and BY individually. Y, RY, and BY for one pixel are each subjected to uniform quantization (linear quantization) PCM with 8 bits. At this time, for example, with respect to the building board 2 having a width of 500 mm and a length of 3000 mm, since the aspect ratio of the image is 3/4, scanning is performed with the short side of the image l = 500 × 3/4 = 375 [mm]. The number of lines is 480, which is 480 for a length of 3000 mm.
× 3000/375 = 3840 [lines]. Therefore, the color image conversion data processing amount is (8 + 8 + 8) ×
640 × 3840 = 59.0 [Mbit] = 7.4 [M
B].

Each of the signals Y, RY, and BY is expressed by the following equation with respect to the three primary color signals of red R, green G, and blue B: Y = 0.3R + 0.59G + 0.11B RY = 0.7R −0.59G−0.11B BY = −0.3R−0.59G + 0.89B

As shown in FIG. 5 (b), in the level conversion, the total amplitude of the quantized signal is set to 2 ⁸ = 256 levels, the luminance signal Y is set with the level 16 as a black level reference and the level 235 as a white level reference. The color signals RY and BY are converted into 220 levels of levels 16 to 235, and the full amplitude of the color signals is converted into 224 levels based on level 128.

FIG. 6 is a block diagram showing the basic configuration of the inspection plate system according to the present invention. An image signal picked up by the TV camera 1 is received by a normal TV receiver 6, projected on a monitor display 7 as a monitoring video, and simultaneously input to an image processing device 8 described later in detail. A signal from a photoelectric switch 9 for detecting the building board 2 is input to the image processing device 8. Further, a keyboard 10, a printer 11, and an image memory 12 are connected in the same manner as a normal computer. The image processing device 8 includes a counter that counts the number of building boards 2, a counter that counts spot-formed pixels, and a counter that counts label numbers.
(Occurrence of chipping, etc.), hue 13b, spot size 13c, spot number 13d and spot distribution 13e are displayed as abnormalities.

FIG. 7 is a block diagram showing the functional configuration of the inspection plate system according to the present invention. The image 20 is scanned 21 by the TV camera 1 and subjected to photoelectric conversion 22 to generate an image signal 2
3 and transmitted to the TV receiver 6, subjected to luminance / color signal separation (Y / C separation) 24, subjected to display processing 25, and displayed on the monitor display 7 as a monitoring image.
The image signal subjected to the luminance / color signal separation 24 is sampled 26, quantized and coded 27, stored as image data 28, and subjected to a plate processing 29.

FIG. 8 shows a storage data structure of pixel values. It is not intended to form digital images,
Two image signals obtained from the TV camera 1 (the luminance signal Y
And two color difference signals RY and BY) are digitally converted, and the image data is stored in a 3840 × 640 matrix array format as described above. Further, the intermediate data d (x _i ,
A storage area of the same 3840 × 640 matrix is prepared for storing y _i ).

FIG. 9 shows an imaging state of an enlarged image of the building board 2 according to another embodiment of the present invention. In the case of inspecting a fine spot pattern such as a sputter-painted product, it is difficult to discriminate a normal image, so an enlarged image of a predetermined magnification is taken. In this case, as shown in FIG. 9, an inspection is performed on a portion having a width m ₁ extending in the longitudinal direction of the building board 2. This makes it possible to inspect the number and size of spots, the scattered state, and the shape of the spot pattern. In addition, by synthesizing each captured image using several TV cameras 1, it is possible to inspect the entire building board 2. Further, a high-definition TV camera may be used as the TV camera 1.

FIG. 10 is a diagram (part 1) for explaining the enlarged image processing in the inspection board system according to the present invention. Consider the case of FIG. In the enlarged captured image l ₁ × m ₁ ,
The superimposed imaging area is an area excluding the movement of the building board within the imaging range.

FIG. 11 is a diagram (part 2) for explaining the enlarged image processing. Consider according to the case of FIG. Building board L × m
₁ minute of the captured image _{P 1 + d 1 + d 2} + rectangular region d ₁ of _{··· + d n, d 2,} ···, each of the first line number x of d _n
As in the case of FIG. 3, x: 480 = (l ₁ −L / 30): l ₁ x = 480 (l ₁ −L / 30) / l ₁ That is, the rectangular area d _1, d _2, · · ·, for capturing images of d _n is, by performing the sampling of the image data to the frame area up to 480 onward from the line corresponding to the x value determined here good.

FIG. 12 is a diagram for explaining the handling of image data classified by the shape characteristic of the surface of the building board 2. A description will be given of the handling of image data for a general building board 2 in which a convex portion of a handle portion is formed as a substantially flat surface and a concave portion (groove portion) is formed and has a real portion. .

The convex, concave (groove) and real parts of the handle are 1
Each of the two building boards 2 forms a clearly distinguishable portion in terms of hue, shape, and the like even when viewed with the naked eye, and presents data values that are clearly distinguished even in image data. Therefore, image evaluation is performed for each of these parts individually.

[In the case of a normal image] For a convex portion, the occurrence of a hue change or a stain is evaluated. (Details will be described later.) The hue change and the generation of dirt are evaluated in the same manner as in the case of the convex portion only when the groove portion cannot be ignored particularly in comparison with the convex portion (for example, when the area is large). Regarding the pattern portion and the real portion, the change in the contour of the peripheral portion (the occurrence of a chipped portion) is evaluated. That is,
It is detected whether or not the set of pixels corresponding to each of the vertical and horizontal straight lines forming the contour of the building board 2 has been interrupted (see FIG. 19).

[In the case of an enlarged image] The number of speckles per predetermined area, the size distribution of speckles, and the scattered state of speckles are evaluated for the convex portions (see FIGS. 22 to 24). Next, an example of handling of image data based on a difference in a convex shape pattern will be described. The evaluation reference value in each of the following cases is set from experience values.

(A) In the case of a single-color configuration, it is possible to detect the occurrence of hue variation and dirt by hue detection. (b) In the case of two colors (ground color + spot color) →
Hue variation is evaluated by obtaining, as a parameter, the ratio between the integrated average value of the ground color forming pixels and the integrated average value of the speckled color forming pixels.

(C) In the case of three colors (ground color + two speckles) → the ratio of the integrated average value of the ground color forming pixels to the integrated average value of the first speckle color forming pixels in a predetermined area, And the ratio of the integral average of the first speckle color forming pixel to the integral average of the second speckle color formation pixel as the three parameters. The evaluation also evaluates the relationship between spots having different hues.

(D) In the case of two colors (ground color + scattered aggregate color)-> For scattered aggregate color, it is often sufficient to evaluate luminance data rather than hue data. For example, when the scattered aggregate is granular, such as sand or metal pieces, a projection is formed on the surface of the building board 2, so it is better to compare the luminance difference rather than the hue difference. Is appropriate. The specific evaluation method conforms to the above method (b).

(E) In the case of three colors (ground color + spot color + scattered aggregate color) → The specific evaluation method is in accordance with the above method (c). (f) In the case of 4 colors (ground color + 2 spots + scattered aggregate color) → For specific evaluation methods, ground color vs. 1st spot color, ground color vs. 2nd spot color, ground color vs. Scatter aggregate color, 1st spot color vs 2nd
Are evaluated as five parameters obtained for the spot color (the ratio is represented by A) and A versus the scattered aggregate color.

(G) 5 colors (ground color + 2 spots + spray aggregate color 2)
In the case of (color) → The specific evaluation method follows the method of (f). A new ratio of the first scattered aggregate color to the second scattered aggregate color (hereinafter referred to as B) and a ratio of A and B are added, and the results are evaluated as seven parameters. (h) In the case of a surface pattern having large irregularities → The color of the reference plate is recorded in advance as comparative data, and the color is measured and compared with the measured color for evaluation.

A more specific procedure of the above evaluation method is shown in FIG.
0 and FIG. The handling of the image data for the hue evaluation of the surface of the building board 2 will be described with reference to FIG. The hue evaluation is performed for each predetermined area _Aij which is a square having a side of a = 125 mm, for example, to evaluate the variation. By the way, the following factors can be considered as factors that determine the state of spots applied to the surface of the building board 2.

(1) Spray conditions, that is, nozzle diameter, nozzle paint outlet shape, spray jet pressure, spray angle, spray distance, number of spray guns used (same color, different color), etc. (2). Reciprocating condition of spray gun, that is, building board 2
Reciprocating reciprocating speed relative to the speed of

(3) The surface unevenness state (shape characteristic) of the building board 2, for example, the surface shape is flat, almost flat (in the case of having fine unevenness like a satin pattern), tile-shaped (by joints or grooves). Some are divided into small sections), others have a staggered tone (when the inclined surface is continuous), and others have a noticeable unevenness. (4). Spraying materials, that is, types of paints (viscosity, particle size, solvent and others) and aggregates (shape, particle size, gloss and others).

(5) In addition to the general coating conditions described in (1) to (4) above, in performing the continuous processing of the building board 2, "the building board runs at high speed". But also has a considerable influence. In other words, the paint in the compression system is released into the air, and is divided into fine droplets while repeatedly colliding with air molecules. Because of this state, the paint that has been turned into droplets is in a state that is susceptible to the effect of airflow generated by the building board 2 that is traveling. Therefore, pixel detection suitable for each speckle state should be performed.

As shown in FIG.
mm * 342 mm, the area of a building board drawn by one pixel as a screen configuration of 640 * 480 pixels is (455 × 34
2) / (640 × 480) ≒ 0.5 [mm ² ],
The length of one side is (0.5) ^0.5 ≒ 0.7 [mm]. Therefore, the detectable spot state is such that the minimum diameter of a spot approximated by a circle is 0.7 [mm], that is, one pixel, and the maximum major axis of a spot approximated by an ellipse is 2.8 [mm], that is, about four pixels. The inspection plate is described on the assumption that the maximum minor axis is 1.4 [mm], that is, about 2 pixels, and that each spot is independently applied so that the number of spots can be counted. .

The speckles are handled as one figure block figure (a figure having a spread) having the same hue. The size of the spot is determined by counting the number of constituent pixels of the detected block figure, and the number of spots is determined by counting the number of detected block figures. In addition, the scattered state of the spots is evaluated based on the representative position information about each detected block figure.

In the above-described pixel detection, it is difficult to detect spots in the case of a normal image, and therefore, data for an enlarged image is actually processed. In this case, the evaluation result is for a portion having a predetermined width (m _{1 in} FIG. 9) in the lateral direction of the building board 2. Where T
By facilitating a plurality of V cameras 1, it is possible to evaluate the entire surface of the building board 2.

As shown in FIG.
If there is no pixel of the same hue at the location, the figure block figure is composed of one pixel. As shown in the figure, in the block diagram composed of a plurality of pixels, the same hue pixel is continuous on the right side (in the scanning line direction) of the target pixel, or the same hue pixel exists in at least one of the positions 2 to 4. Can be recognized by detecting that

A specific example of the detection of the block figure will be described for an enlarged image with reference to FIG. When the individual spots are formed of the same hue paint, the same hue pixels on each line are detected from left to right, and when the same hue pixels are consecutive, the same number is assigned (for example, the scanning line L1). 2 or 3),
The number is recorded (for example, 2 of the scanning line L1 becomes 3
Three and two). If not consecutive, set the number to 1.
(For example, it is increased to 5 in the scanning line L1). This operation is performed step by step on the lower lines in order from the uppermost line. At this time, in the detection of the second and lower pixels, one of the pixels or one of the consecutive pixels is 1
It is checked whether or not it is adjacent to the same hue pixel existing in the next upper row, and if it is adjacent, the same number is assigned (for example, 2 or 3 of the scanning line L2). If they are not adjacent to each other, the number is further increased by one (for example, 6 of the scanning line L3).

As described above, the final additional number (9 in the figure) as a result of performing the pixel detection on the image data corresponding to the inspection target area indicates the number of spots.
A value obtained by adding the number of pixels of the same additional number recorded for each line (for example, 7 for 2) indicates the size of each spot.

In the case of a pattern having spots of different hues,
Numbering is performed for each spot. The evaluation of the scattered state of speckles will be described with reference to FIG. Since the sputter is applied when the spray gun traverses the running building board 2 in the short direction, the rightmost point or the leftmost point (FIG. (Pixel surrounded by a bold frame) is a point where the paint first arrives at the building board 2. Then, a spot-like figure block figure is formed starting from the point. Therefore, the scattered state of the speckles can be evaluated by specifying the right end point or the left end point in each spectacle as a parameter representing each speckle position related to the sputtering direction.

The scattered pixels (position information) shown in FIG. 16 have a correlation with the movement locus represented by the spray gun shown in FIG. That is, for the locus of
The spots applied by spattering on the building board 2 belonging to the range that has been moved obliquely downward to the left with the width p by the multiple spray guns are involved. Spots belonging to the moved range are relevant. However, a spot pattern in a substantially uniform scattered state is formed by overpainting as shown in FIG.

Next, a main flow of the image processing apparatus will be described with reference to FIG. First, it is determined whether or not the system power is on (S1). If NO, the process waits until the system power is turned on. If YES, the corresponding condition value is read from a file stored in advance to perform initialization (S1).
2). In step S3, it is determined whether or not the imaging screen is a normal image (that is, an enlarged image). If YES, a condition setting for determining a sampling condition is performed (S4).
It is determined whether or not the building board 2 has been detected (S5). If NO, the process waits. If the building board 2 is detected and the result is YES, data sampling is started (S6) and the number of the building boards 2 is counted. The counter to be executed is increased by 1 (S7). Then, the sampling data is processed and accumulated as described above (S8), and first, a board inspection (I) of the outline shape of the building board is performed (S9). That is, H1, H2, V1, V2, C shown in FIG.
The conditions for specifying data relating to 1 and C2 are set (S101), the straight line portion of H1 is evaluated in step S102, the straight line portion of V1 is evaluated in step S103, and the corner portion of C2 is evaluated in step S104. Step S105
Evaluates the linear portion of H2, evaluates the linear portion of V2 in step S106, and evaluates the corner of C1 in step S107. As described above, in steps S102 to S107, the corresponding image data group is evaluated as abnormal when the luminance value exceeds the allowable range. And step S
At 108, it is determined whether or not there is an abnormality in each part. If the answer is NO, the process returns to step S10. If the answer is YES, an abnormality indicator lamp for indicating that the contour is abnormal is turned on ( S109), and waits for the end of the corresponding action (S1)
10) When the countermeasure such as removing the target building board 2 is completed, the abnormality indicator lamp is turned off (S111).

Next, the inspection board (II) for the hue of the building board is implemented (S10). As shown in FIG. 20, first, step S
In 201, the image data of the groove portion unnecessary for the hue evaluation is excluded,
Setting of constant values and initialization of evaluation area parameters (i ← 1,
j ← 1) and the like, and it is determined in step S202 which pattern of the pattern shape belongs to the pattern shown in FIG. 12. If it is the pattern (a), the process proceeds to step S203, and the evaluation area is determined. Of the extreme values D of the signals Y, RY, and BY are excluded, and the remaining signals Y, RY, and B-
The average value of each Y is calculated (S204), and 2
The square root is calculated and used as the provisional hue value CR1 of the ground color (S20
5) If this value is within the allowable range (C ₁ + α ≧ CR1 ≧
It is determined whether or not (C ₁ -α) is satisfied (S206). If YES, the CR1 value is recorded (S207), the process moves to the next evaluation area (S208), and it is determined whether all the evaluation areas have been evaluated. Judgment is made (S209), and if all have not been exhausted (NO), the same evaluation is performed in the next evaluation area. If the evaluation is completed for all the evaluation areas (YES) C
Statistical evaluation of the variation in the R1 value is performed (S210), and the evaluation result is output to a printer (S211). If CR1 is not within the allowable range in step S206 (in the case of NO), the abnormality display lamp for indicating that the hue is abnormal is turned on (S212), and the end of the corresponding action is waited for (S21).
3) When the abnormality processing is completed, the abnormality indicator lamp is turned off (S214).

If it is determined in step S202 that the pattern is the pattern (b), the basic procedure is the same as that of the pattern (a) as shown in FIG. 21, but the ground color hue value CR1 and the spot hue value CR2 is calculated (S215), and the ratio CR2 / C is calculated.
R1 is calculated (S216), and it is determined whether or not the ratio is within an allowable range (C ₂ + β ≧ CR2 / CR1 ≧ C ₂ −β) (S217). If YES, the ratio is recorded (S217). S21
8) Thereafter, the same processing (S219) as for pattern (a) is performed, and if NO, the same processing (S219) as for pattern (a) is performed.
220).

When it is determined in step S202 that the pattern is the pattern (c), as shown in FIG.
The hue value CR2 of the first spot and the hue value CR3 of the second spot are calculated (S221), and the ratios CR2 / CR1, CR3 /
The CR1 and CR3 / CR2 calculated (S222), as the ratio is within the allowable range _{(C 3 + γ ≧ CR2 /} CR1 ≧ C 3
-Γ, C ₄ + δ ≧ CR3 / CR1 ≧ C ₄ −δ and C ₅ + ε
_{≧ CR3 / CR2 ≧ C 5 -ε} ) whether the determined (S22
3) If YES, the ratio is recorded (S224), and the same processing (S225) as for pattern (a) is performed. If NO, the same processing as for pattern (a) is also performed (S226).

For pattern (d), Ave.Y1 (average of aggregate luminance data) is used instead of CR2 of pattern (b) (not shown). For pattern (e), Ave.Y1 is used instead of CR3 of pattern (c) (not shown).

If it is determined in step S202 that the pattern is the pattern (f), as shown in FIG.
The hue value CR2 of the first spot, the hue value CR3 of the second spot, and the average value Ave.Y1 of the aggregate luminance data are calculated (S22).
7), the ratios CR2 / CR1, CR3 / CR1, Ave.
Y1 / CR1, CR3 / CR2 and Ave.Y1 · CR2
/ CR3 is calculated (S228), and their data is managed (S229).

If it is determined in step S202 that the pattern is (g), the hue value CR1 of the ground color, the hue value CR2 of the first spot, the hue value CR3 of the second spot, and the average value of the first aggregate luminance data Ave. Average value Av of Y1 and second aggregate luminance data Av
e. Calculate Y2 (S230) and calculate its ratio CR2 / CR1,
CR3 / CR1, Ave.Y1 / CR1, CR3 / CR
2, Ave.Y1 / CR2 / CR3, Ave.Y2 / Ave.Y
Y1 and CR2.Ave.Y2 / CR3.Ave.Y1 are calculated (S231), and their data is managed (S23).
2).

If it is determined in step S202 that the pattern is (h), the color CRS of the reference plate is recorded in advance as comparison data, and the color CRS is evaluated by comparing and managing the measured color CRM. (S233). here,
Returning to FIG. 18, at step S11, it is determined whether or not the number of inspection plates has reached the set number n. If NO, the process returns to step S5, and if YES, the inspection plate ends.

If it is determined in step S3 that the image is not a normal image, that is, if the image is an enlarged image, steps S12 to S16 are performed in the same manner as steps S4 to S8, and then a spot pattern inspection board (III) is executed (S17). ). As shown in FIG. 24, first, in step S301, conditions such as initialization of evaluation area parameters (i ← 1, j ← 1) are set, and in step S302, the number of spot colors is determined. , Searching for spot forming pixels, labeling of the searched pixel,
The number of spots in a predetermined area, the spot size status,
The detection (IV) of the scattered spot state is performed (S303, S3)
04, S305).

More specifically, referring to FIG. 25, first, the label counter value for distinguishing each spot and counting the number of spots is 1, and the number counter value which is the number of pixels belonging to each spot and is a parameter of the spot size. Is set to 0 (S4
01), the row number i is set to 1 (S402), i = 480
The process enters a spot formation pixel search loop up to (S403).
The column number j is set to 1 (S404), and a spot forming pixel search loop up to j = 640 is entered (S405).
It is determined whether or not> 1 (S406), and if NO and the first line, it is determined whether or not the pixel is of a spot color (S407). The number counter value is incremented by one (S408), and a label counter value is labeled for the detected pixel (S409). Furthermore, j
The next pixel is set as ← j + 1 (S410), and it is determined whether or not the pixel is a spot color (S411). If = is the spot color, the number counter value is incremented by one (S41).
2) Label the detected pixel with a label counter value (S413). In step S411, it becomes ≠, and if the pixel is not a spot color pixel, the number counter value is recorded in association with the label counter value (S414), the label counter value is incremented by 1 (S415), and the number counter is cleared (S416). . If YES in step S406 and the second and subsequent rows, as shown in FIG. 26, it is determined whether or not the pixel is a spot color (S417). The value is incremented by 1 (S4
18) It is determined whether there is connectivity with the upper pixel (S419), and if so, the detected pixel is labeled with the number of the upper pixel (S420). If not, the detected pixel is labeled with the label counter value as a temporary label (S421). Further, as j ← j + 1, the process moves to the next pixel (S422), and it is determined whether or not the pixel is a spot color (S423). (S424), it is determined whether or not there is connectivity with the upper pixel (S425), and if so, the temporary label is changed to the number of the upper pixel for the detected pixel (S426). If not, a label counter value is labeled for the detected pixel (S427). In step S423, it becomes ≠, and if it is not a pixel of a spot color, the number counter value is recorded in association with the label counter value (S428), and the label counter value is incremented by one (S42).
9), the number counter is cleared (S430). In step S407 in FIG. 25, the result is ≠, and if it is not a pixel of a spot color, j ← j + 1 is set, and the process moves to the next pixel (S43).
1), exits the loop when j = 640 (S4
32), as i ← i + 1, the process moves to the next scanning line (S43)
3) Exit the loop when i = 480 (S4
34).

Thereafter, the label counter value is recorded in FIG. 27 (S435), the label counter is cleared (S436), the recorded number of connected pixels for each label number is added, and the added value is recorded. (S437), and the variation of the added value is evaluated (S438). If the speckle scatter is not to be inspected (NO in S439), the process proceeds to step S30.
6 to check for scattered spots (YE in S439).
S), the rightmost point of the top row (corresponding to the bold pixel in FIG. 16) is specified and recorded in the pixel set for each label (S44).
0), the uppermost left end point in the pixel set for each label is specified and recorded (S441). Then, the distance (the number of pixels between pixels) l _RH to the pixel on the right side in the horizontal direction of each pixel (if there is no pixel) and the distance l to the vertically lower pixel (if there is no pixel) _RD is obtained and recorded (S442), and the left end point set pixel is obtained and recorded as L _LH and L _LD similarly to step S442 (S443). Then, the dispersion is evaluated for each of l _RH , l _RD , l _LH , and l _LD (S444) and step S306.
Return to

Returning to FIG. 24, if the number of speckles, the size of speckles, and the status of speckles are not NO in step S306 compared with the pre-recorded data, the detection result is output to the printer. Do (S310)
Returning to step S18, if there is an abnormality in the case of YES, the respective abnormality display lamps indicating the abnormality are lit (S18).
307), wait for the end of the corresponding action (S308), and after the end of the corresponding action, turn off the abnormality display lamp (S309).
It returns to step S18.

If the number of spot colors is two in step S302, the process for one color is individually executed for each color, and the respective result data are evaluated together (S311). Returning to FIG. 18 again, it is determined in step S18 whether or not the number of inspection plates has reached the set number n. If NO, the process returns to step S13, and if YES, the inspection plate ends. This is the end of the description of the main flow of the image processing apparatus.

[0059]

Therefore, according to the present invention, (1) there is no need to stop the line for the inspection plate, and the inspection plate can be online. Accompanying this, by removing the found defective plate by an appropriate method, it is possible to prevent the defective plate from being erroneously shipped.

(2) It is possible to perform a highly accurate inspection board even when a finely patterned board such as a sputter coating is targeted. Specifically, it is possible to measure the size and number of speckles, and to determine the scattered state of speckles.

(3) An ordinary TV camera imaging system can be used without any special equipment. Therefore, no new equipment installation space is required and no increase in equipment cost is caused, so that the introduction of the line is not hindered. (4) By removing any of the signals obtained by imaging the same part of the building board with different frames, the amount of image signals to be processed can be reduced. (5). By removing the groove data and inspecting the hue of the building board, it is possible to inspect the coating state of the building board purely by inspecting the hue of only the convex part .

[Brief description of the drawings]

FIG. 1 is a perspective view showing an imaging state of a building board by a TV camera in the present invention.

FIG. 2 is an explanatory diagram (part 1) of the normal video processing of the present invention.

FIG. 3 is an explanatory diagram (part 2) of the normal video processing of the present invention.

FIG. 4 is an explanatory diagram (part 3) of the normal video processing of the present invention.

FIG. 5 is a diagram illustrating sampling and encoding of image data according to the present invention.

FIG. 6 is a block diagram showing a basic configuration of the present invention.

FIG. 7 is a block diagram showing a functional configuration of the present invention.

FIG. 8 is a diagram showing a storage data structure of a pixel value in the present invention.

FIG. 9 is a perspective view showing an imaging state of an enlarged image of a building board by a TV camera in the present invention.

FIG. 10 is an explanatory diagram (part 1) of the enlarged video processing of the present invention.

FIG. 11 is an explanatory view (part 2) of the enlarged video processing of the present invention.

FIG. 12 is an explanatory diagram of handling of image data classified by shape characteristic of a building board surface portion.

FIG. 13 is an explanatory diagram of handling of image data for hue evaluation of a building board surface portion.

FIG. 14 is an explanatory diagram (part 1) of a basic analysis of image data.

FIG. 15 is an explanatory diagram (part 2) of the basic analysis of image data.

FIG. 16 is an explanatory diagram (part 3) of a basic analysis of image data.

FIG. 17 is an explanatory diagram (part 4) of a basic analysis of image data.

FIG. 18 is an explanatory diagram of a main flow of the image processing apparatus.

FIG. 19 is an explanatory diagram of a contour shape inspection flow of the image processing apparatus.

FIG. 20 is an explanatory diagram (part 1) of a hue inspection flow of the image processing apparatus.

FIG. 21 is an explanatory diagram (part 2) of the hue inspection flow of the image processing apparatus.

FIG. 22 is an explanatory diagram (part 3) of the hue inspection flow of the image processing apparatus.

FIG. 23 is an explanatory diagram (part 4) of the hue inspection flow of the image processing apparatus.

FIG. 24 is an explanatory diagram of a spot pattern inspection flow of the image processing apparatus.

FIG. 25 is a detailed explanatory diagram (part 1) of the inspection pattern inspection flow of the image processing apparatus.

FIG. 26 is a detailed explanatory diagram (part 2) of the spot pattern inspection flow of the image processing apparatus.

FIG. 27 is a detailed explanatory diagram (part 3) of the spot pattern inspection flow of the image processing apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 TV camera 2 Building board 3 Imaging screen 4 Guide 6 TV receiver 7 Monitor display 8 Image processing device 9 Photoelectric switch 10 Keyboard 11 Printer 12 Image memory 13 Abnormal display lamp group 20 Image 21 Scanning 22 Photoelectric conversion 23 Image signal 24 Y / C separation 25 Display processing 26 Sampling 27 Quantization and coding 28 Image data storage 29 Inspection processing

──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. ⁷ , DB name) G01N 21/86 G01N 21/892 G01B 11/24 G01B 11/30 G01J 3/46

Claims (7)

(57) [Claims] 1. A board inspection system for inspecting a building board, wherein the TV camera captures an image of the running building board;
An inspection board system comprising: an image processing unit that processes an image signal by removing any one of signals obtained by imaging the same part of the building board in a different frame from image signals from a camera. 2. The plate detecting system according to claim 1, wherein said image processing means performs predetermined image signal processing during a time when said removed signal is present. 3. The inspection board system according to claim 1, wherein the image processing unit inspects a contour shape of the building board. 4. The image processing device according to claim 1, wherein the image processing means detects spots on the building board.
Inspection plate system as described. 5. The image processing device according to claim 1, wherein the number of the spots is:
The inspection plate system according to claim 4, wherein the size of the spots or the scattered state of the spots is evaluated. 6. A board inspection system for inspecting a building board, wherein the TV camera captures an image of the running building board;
An image processing means for processing an image signal obtained by removing groove data from an image signal from a camera to inspect a hue of the building board. Wherein said TV camera, inspection panel system according to any one of claims 1 to 6, characterized in that it can change the imaging magnification.