JP2022089476A - Device and method for discriminating concave-convex defects in workpiece, and program - Google Patents

Abstract

To provide a device and a method for discriminating concave-convex defects, as well as a program that can discriminate whether a surface defect in a workpiece is a concave defect or a convex defect regardless of the shape of a measurement part, while at least one of the workpiece, illumination, and photographing means is made to relatively move by using only the illumination from one direction.SOLUTION: Photographing means 2 performs photographing while at least one of a workpiece 100, illumination light, and the photographing means is made to relatively move. Based on a plurality of images 20 including an image having a bright part 21 and images having dark parts 22, 23 obtained through the photographing, an approximate expression is calculated, the approximate expression representing the relationship between the position of a dark point 31 appearing in the bright part 21 as a surface defect in the workpiece, the positions of bright points 32 appearing in the dark parts 22, 23 as surface defects in the workpiece, an index indicating the relative position of the workpiece 100 to the illumination light or the photographing means 2, and the position(s) of either one or both of the bright points and the dark point. A concave-convex discrimination index indicating the relative positional relationship between the dark point 31 and the bright points 32 is calculated from the approximate expression, and whether the surface defect is concave or convex is determined based on the concave-convex discrimination index.SELECTED DRAWING: Figure 9

Description

The present invention relates to a defect unevenness discriminating device for a work, and a method and a program for discriminating whether a surface defect occurring on a painted surface or the like of a work such as a vehicle body is a concave defect or a convex defect.

For example, it has been conventionally performed to detect a surface defect on a painted surface of a vehicle body and repair the defective portion, but the method of repairing the defect differs depending on whether the surface defect is a concave defect or a convex defect. Therefore, it is required not only to detect surface defects but also to determine whether the surface defects are concave defects or convex defects.

As a method for discriminating whether the surface defect of such a work is a concave defect or a convex defect, Patent Document 1 is a surface defect detecting method for optically detecting a surface defect of a steel material, and can discriminate between two or more. Using a light source, the same inspection target site is continuously irradiated with linear illumination light in a direction orthogonal to the moving direction of the steel material from different directions at substantially the same incident angle to acquire an image by reflected light, and between the acquired images. The surface defects in the inspection target part are detected by performing the difference processing in, and the adjacent bright and dark parts of the image obtained by performing the difference processing between the acquired images are extracted and adjacent to each other. A method for determining the presence or absence of uneven surface defects from the positional relationship between the bright and dark areas and the irradiation direction of the illumination light is disclosed.

Further, in Patent Document 2, a flat plate is irradiated with light and dark illuminations in parallel, the boundary between light and dark is observed with a line sensor, and one-dimensional images taken by the line sensor are arranged in order while moving the plate. A technique is disclosed in which a two-dimensional image is synthesized, an image in which a bright blob and a dark blob are adjacent to each other in a concave defect portion or a convex defect portion is obtained, and whether it is a concave defect or a convex defect is determined based on the positional relationship between the light and dark. ..

Japanese Patent No. 6394514 US Pat. No. 7,105,848

However, the technique described in Patent Document 1 has a problem that the apparatus becomes complicated because it is necessary to use two or more distinguishable light sources.

Further, in the technique described in Patent Document 2, when the work is reflected and measured by a line sensor, even if the work is a flat surface, undulations occur depending on the measurement site, or the work is a three-dimensional free curved surface. There is a problem that the defect itself cannot be measured.

The present invention has been made in view of such a technical background, using only illumination from one direction and measuring while relatively moving at least one of the workpiece or illumination or imaging means. It is an object of the present invention to provide a work defect unevenness discrimination device capable of discriminating whether a work surface defect is a concave defect or a convex defect regardless of the shape of a portion, the same method, and a program.

The above object is achieved by the following means.
(1) A lighting device that illuminates the work with illumination light,
Shooting means and
A means of transportation that moves at least one of the work or illumination light or the means of photography relative to the other.
Equipped with
The work illuminated by the illumination light is photographed by the photographing means while the work or the illumination light or at least one of the photographing means is relatively moved relative to the other by the moving means, and further.
Based on a plurality of images obtained by the imaging, including an image having a bright portion corresponding to an irradiated portion of the illumination light or an image having a dark portion corresponding to a non-irradiated portion of the illumination light, the surface of the work is formed on the bright portion. Both the position of the dark spot appearing as a defect, the position of the bright spot appearing as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means, and either the bright spot or the dark spot. A calculation method for calculating an approximate expression that expresses the relationship with the position,
From the approximate expression calculated by the calculation means, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and the discrimination means for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A device for determining the defect unevenness of a work, which is characterized by being equipped with.
(2) The defect unevenness discrimination device for a work according to item 1 above, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting.
(3) The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point have a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. The defect unevenness discrimination device for the work according to the above item 1 or 2.
(4) The defect unevenness discrimination apparatus for a work according to any one of the above items 1 to 3, wherein the sign of the unevenness discrimination index is changed according to the relative moving direction of the work with respect to the photographing means or the illumination light.
(5) The defect unevenness determination device for a work according to any one of 1 to 4 above, which rotates the coordinates of dark points and bright points on an image or the image according to the rotation angle of the photographing means with respect to the optical axis of the photographing means. ..
(6) As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, and the coordinates of the illumination light when the illumination light is moved. The device for determining defect unevenness of a work according to any one of 1 to 5 above, which includes at least one of the coordinates of the photographing means when the photographing means is moved.
(7) By the moving means, the work illuminated by the illumination light is moved by the photographing means while at least one of the illumination light or the photographing means for illuminating the work is relatively moved with respect to the other. The shooting steps to shoot and
Based on a plurality of images obtained by the photographing step, including an image having a bright portion corresponding to an illuminated portion of the illumination light and an image having a dark portion corresponding to a non-irradiated portion of the illumination light, the work is formed in the bright portion. Both the position of the dark spot appearing as a surface defect, the position of the bright spot appearing as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means, and either the bright spot or the dark spot. A calculation step to calculate an approximate expression that expresses the relationship with the position of
From the approximate expression calculated by the calculation step, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and a discrimination step for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A method for discriminating defects and unevenness of a work, which is characterized by being provided with.
(8) The defect unevenness discrimination method for a work according to item 7 above, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting.
(9) The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point have a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. The method for determining the defect unevenness of the work according to the above item 7 or 8.
(10) The method for determining defect unevenness of a work according to any one of items 7 to 9 above, wherein the sign of the unevenness discrimination index is changed according to the relative moving direction of the work with respect to the photographing means or the illumination light.
(11) The method for determining defect unevenness of a work according to any one of items 7 to 10 above, wherein the coordinates of dark and bright points on the image or the image is rotated according to the rotation angle of the photographing means with respect to the optical axis of the photographing means. ..
(12) As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, and the coordinates of the illumination light when the illumination light is moved. The method for determining defect unevenness of a work according to any one of items 7 to 11 above, which includes at least one of the coordinates of the photographing means when the photographing means is moved.
(13) By the moving means, the work illuminated by the illumination light is moved by the photographing means while at least one of the illumination light or the photographing means for illuminating the work is relatively moved with respect to the other. Based on a plurality of images obtained by photographing, including an image having a bright portion corresponding to an irradiated portion of illumination light or an image having a dark portion corresponding to a non-irradiated portion of illumination light, the work is formed in the bright portion. Both the position of the dark spot appearing as a surface defect, the position of the bright spot appearing as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means, and either the bright spot or the dark spot. A calculation step that calculates an approximate expression that expresses the relationship with the position of
From the approximate expression calculated by the calculation step, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and a discrimination step for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A program that lets your computer run.
(14) The program according to item 13 above, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting.
(15) The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point have a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. The program according to the preceding paragraph 13 or 14.
(16) The program according to any one of the preceding paragraphs 13 to 15, which changes the sign of the unevenness discrimination index according to the relative moving direction of the work with respect to the photographing means or the illumination light.
(17) The program according to any one of 13 to 16 above, which rotates the coordinates of dark and bright points on an image or the image according to the rotation angle of the photographing means with respect to the optical axis of the photographing means.
(18) As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, and the coordinates of the illumination light when the illumination light is moved. , The program according to any one of 13 to 17 in the preceding paragraph, which includes at least one of the coordinates of the photographing means when the photographing means is moved.

According to the inventions described in the preceding paragraphs (1), (7) and (13), at least one of the work or the illumination light for illuminating the work or the photographing means is relatively relative to the other by means of transportation. Darkness that appears as a surface defect of the work in the bright part based on multiple images including an image having a bright part corresponding to the illuminated part of the illumination light or an image having a dark part corresponding to the non-irradiated part of the illumination light while moving. An approximate expression is calculated that expresses the relationship between the position of the point, the position of the bright point that appears as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means. Then, from the calculated approximate expression, an unevenness discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and the unevenness of the surface defect is discriminated based on the unevenness discrimination index.

In this way, shooting can be performed while moving at least one of the work, the illumination light, and the shooting means relative to the other, and when shooting while moving the work, the work is relative to each other. In addition to the fact that it is not necessary to stop the movement and take a picture, a plurality of lighting devices for illuminating the work from a plurality of directions are not required, and one lighting device is sufficient, and the configuration is simplified. Moreover, since the work is not reflected and measured by the line sensor, it is possible to determine whether the surface defect is a concave defect or a convex defect regardless of the surface shape of the measurement site.

According to the inventions described in the preceding paragraphs (2), (8) and (14), the unevenness discrimination index including the number of times that the dark spot and the bright spot are in a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting. Therefore, simple and accurate discrimination can be performed.

According to the inventions described in the preceding paragraphs (3), (9) and (15), the dark spot or the bright spot when the dark spot and the bright spot are in a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. Since it includes an index showing the certainty of the bright point, it is possible to make a more accurate determination.

According to the inventions described in the preceding paragraphs (4), (10) and (16), even if the relative moving direction of the work with respect to the photographing means or the illumination light changes, accurate discrimination can be made by changing the sign of the unevenness discrimination index. It can be performed.

According to the inventions described in the preceding paragraphs (5), (11) and (17), even if the photographing means is rotated with respect to its own optical axis, dark spots and bright spots on the image are obtained according to the rotation angle. Accurate discrimination can be made by rotating the coordinates or image of.

According to the inventions described in the preceding paragraphs (6), (12) and (18), as an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, and the rotation speed of the rotary encoder provided in the moving means. Since it includes at least one of the measurement time, the coordinates of the illumination light when moving the illumination light, and the coordinates of the photographing means when moving the photographing means, the position of the dark spot appearing in the bright part of the image and the dark part of the image It is easy to calculate an approximate expression that expresses the relationship between the position of the bright spot that appears and the index that indicates the relative position of the work with respect to the illumination light.

It is a block diagram which shows the structural example of the defect unevenness discriminating apparatus of the work which concerns on one Embodiment of this invention. It is a figure which shows an example of the image photographed by the photographing means. (A) to (C) are images taken by the photographing means (the figure on the left side of each figure) and the optical arrangement (the figure on the right side of each figure) when the work having a convex defect moves from left to right. ). (D) to (F) are continuations of FIGS. 3 (A) to 3 (C). (A) to (C) are images taken by the photographing means (the figure on the left side of each figure) and the optical arrangement (the figure on the right side of each figure) when the work having the concave defect moves from left to right. ). (D) to (F) are continuations of FIGS. 5A to 5C. For convex defects, the relationship between each shooting frame number and the X coordinate of the dark point that appears dark on the bright part of the image is represented by a black circle, and the relationship between the shooting frame number and the X coordinate of the bright point that appears bright on the dark part is represented by a white circle. It is a graph that represents and plots these. (A) to (F) are diagrams for explaining a state in which an image is cut out with a rectangular frame from each of the images of FIGS. 3 (A) to (C) and FIGS. 4 (D) to (F). In the images cut out in FIGS. 8A to 8F, the X coordinate of the dark point appearing on the bright part and the X coordinate of the bright point appearing on the dark part were obtained and plotted against the shooting frame number. It is a graph. (A) to (F) are diagrams for explaining a state in which an image is cut out with a rectangular frame from each of the images of FIGS. 5 (A) to (C) and FIGS. 6 (D) to (F). In the images cut out in FIGS. 8A to 8F, the X coordinate of the dark point appearing on the bright part and the X coordinate of the bright point appearing on the dark part were obtained and plotted against the shooting frame number. It is a graph. It is the figure which attached the frame number to the graph of FIG. It is a graph corresponding to FIG. 9, and is the X coordinate of the dark point appearing on the bright part of the image when the work having the convex defect moves from the right direction to the left direction of FIG. 1, and the bright point appearing on the dark part. It is a graph which obtained each X coordinate and plotted against the shooting frame number. It is a graph corresponding to FIG. 11, and is the X coordinate of the dark point appearing on the bright part of the image when the work having the concave defect moves from the right direction to the left direction of FIG. 1, and the bright point appearing on the dark part. It is a graph which obtained each X coordinate and plotted against the shooting frame number. It is a graph which shows the relationship between the coordinates of a dark point and a bright point, and a frame number, which is obtained when the position of a work having a convex defect is fixed and the illumination light is moved. It is a graph which shows the relationship between the coordinates of a dark point and a bright point, and a frame number, which is obtained when the position of a work having a concave defect is fixed and the illumination light is moved.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing a configuration example of a defect unevenness discriminating device for a work according to an embodiment of the present invention.

The defect unevenness determination device of this work includes a lighting device 1, a camera 2 as a photographing means, a calculation unit 3, a display unit 4, and the like.

The lighting device 1 uses a belt conveyor (not shown) as a means of transportation to illuminate the work 100 moving in the direction of arrow A in FIG. 1 on the production line with a single striped illumination light. be. Every time the work 100 moves a certain distance, the work is illuminated by the lighting device 1 and the light reflected by the work 100 is sequentially photographed by the camera 2.

The calculation unit 3 detects surface defects of the work from the image taken by the camera 2 and determines whether the detected surface defects are concave defects or convex defects. The calculation unit 3 has a CPU 3a that performs surface defect detection processing, concave defect or convex defect discrimination processing, other calculations, and control, a ROM 3b that holds an operation program of the CPU 3a and various data, and a CPU 3a according to the operation program. It includes a RAM 3c, which is a work area for executing processing, and a storage unit (not shown) of a hard disk device or the like. The arithmetic unit 3 may be a dedicated device or may be configured by a general-purpose personal computer.

The display unit 4 displays the determination result of whether the surface defect executed by the calculation unit 3 is a concave defect or a convex defect.

In this embodiment, consider a work 100 whose surface is specularly reflected, such as a work surface having a clear coating such as a vehicle body surface or having a metallic luster.

When there is no defect on the surface of the work 100, the illumination light from the lighting device 1 is specularly reflected on the surface of the work 100, and then the image is taken by the camera 2. From the camera 2, a photographed image 20 having a striped pattern in which dark band portions 22 and 23 are present on both sides of the bright band portion 21 as shown in FIG. 2 can be obtained. The bright band portion 21 in the captured image 20 corresponds to the striped light band of the lighting device 1, in other words, corresponds to the irradiation portion of the illumination light. As shown in FIG. 1, the dark band portions 22 and 23 correspond to the regions 11 and 12 where the light bands on both sides of the lighting device 1 do not exist, in other words, correspond to the non-irradiated portion of the illumination light.

However, the area near the boundary between the dark zone portions 22 and 23 in the bright zone portion 21 appears to be slightly darker than the central portion of the bright zone portion 21.

In this embodiment, the lighting device 1 having one light emitting region is used, but for example, a lighting device in which a plurality of striped light emitting parts and non-light emitting parts are alternately displayed by using a liquid crystal display or the like is used. You may.

The position of the work 100 is detected by, for example, an encoder mounted on a belt conveyor on which the work 100 is mounted. Each time the work 100 moves by a certain amount, a trigger is generated from a signal from the encoder to take a picture, whereby the position of the image on the surface of the work 100 is specified.

In the example of FIG. 1, the flat plate work 100 is shown for simplicity, but the bright band portion 21 due to the illumination light and at least one dark band portion 22 and 23 adjacent to the bright band portion 21 are captured by the camera 2. As long as the work 100 has an inclination within the included range, a free curved surface such as a vehicle body may be used.

Next, when the work 100 having a surface defect moves from left to right as shown by the arrow A in FIG. 1, how the surface defect appears in the image 20 captured by the camera 2 will be described.

When a surface defect is present, the reflection direction of the illumination light changes on the defect surface, so that light from the illumination device 1 at a different position is incident on the photographing surface of the camera 2 as compared with the case where the defect is not present.

Therefore, when there is a defect on the surface of the work 100, an image having a brightness different from that around the defect can be obtained at the defective portion.
<In the case of convex defects>
In FIGS. 3A to 3C and FIGS. 4D to 4F, images 20 taken by the camera 2 when the work 100 having the convex defect 110 moves from left to right (in each figure). The left side figure) and the optical arrangement at that time (the right side figure of each figure) are shown. In each of the captured images 20 of FIGS. 3 (A) to 3 (C) and FIGS. 4 (D) to 4 (F), the dark band portions 22 and 23 are also shown in a white state like the bright band portion 21 for convenience. However, it actually appears dark as shown in Fig. 2. The same applies to FIGS. 5 and later.

3 (A) to 3 (C) show the captured image 20 and the optical arrangement when the position of the convex defect 110 on the camera photographing surface is near the position where the left edge portion of the light band of the lighting device 1 is reflected. It is shown.

In the measurement systems of FIGS. 3A to 3C, the light emitted from the right illumination position 1a is reflected on the right slope of the convex defect 110 and incident on the camera 2 as compared with the case where the convex defect 110 does not exist. do. On the other hand, as compared with the case where the convex defect 110 does not exist, the light emitted from the illumination position 1b on the left side is reflected on the left slope of the convex defect 110 and then incident on the camera 2.

As a result, as shown in the left figures of FIGS. 3A to 3C, the left slope of the convex defect 110 is darkly reflected as a dark point 31, and the right slope is brightly reflected as a bright point 32. In FIGS. 3A to 3C and the drawings after FIG. 4, the dark spot 31 is indicated by a black circle and the bright spot 32 is indicated by a white circle.

When there is a convex defect 110 at the position of FIG. 3A, the left slope of the convex defect 110 has the same amount of light as the dark band portion 22 around the defect on the shooting surface of the camera 2, so that it is a dark spot. 31 is hidden behind the dark zone 22 and cannot be seen. On the other hand, since the right slope of the convex defect 110 is brighter on the shooting surface of the camera 2 than the dark zone portion 22 around the defect, the bright point 32 appears to shine brightly.

When there is a convex defect 110 at the position of FIG. 3B, the left slope of the convex defect 110 comes to the position where the bright band portion 21 is reflected, so that the dark spot 31 appears darker than the periphery of the defect. On the other hand, the right slope of the convex defect 110 reflects the light emitted from the brightest part in the central part of the illumination, while the image of the left end portion of the light band is reflected around the defect on the photographing surface, and the central part of the optical band. It is a little darker than. Therefore, on the photographing surface, the right slope of the convex defect 110 appears to shine brighter than the periphery as a bright point 32.

When there is a convex defect 110 at the position of FIG. 3C, the right slope of the convex defect 110 has the same amount of light as the bright band portion 21 around the defect on the shooting surface of the camera 2, so that the bright point 32 is It is hidden behind the bright band 21 and cannot be seen. On the other hand, since the left slope of the convex defect 110 is darker on the shooting surface of the camera 2 than the portion where the surrounding light band is reflected, the dark spot 31 appears dark.

4 (D) to 4 (F) show the captured image 20 and the optical arrangement when the position of the convex defect 110 on the camera photographing surface is near the position where the right edge portion of the light band of the lighting device 1 is reflected. Has been done.

Contrary to FIGS. 3A to 3C, the left slope of the convex defect 110 appears bright and the right slope appears dark. The appearance on the camera shooting surface is the same. In FIG. 4D, the bright point 32 corresponding to the left slope is hidden by the surrounding bright band portion 21, and the dark point 31 corresponding to the right slope is compared with the surroundings. It looks dark.

In FIG. 4E, the bright point 32 corresponding to the left slope of the convex defect 110 appears brighter than the surroundings, and the dark point 31 corresponding to the right slope appears dark.

In FIG. 4 (F), the dark point 31 corresponding to the right slope of the convex defect 110 is hidden by the surrounding dark zone portion 23, and the bright point 32 corresponding to the left slope appears bright.
<In the case of concave defect>
In FIGS. 5A to 5C and FIGS. 6D to 6F, images taken by the camera 2 when the work 100 having the concave defect 120 moves from left to right (left side of each figure). The figure) and the optical arrangement at that time (the figure on the right side of each figure) are shown.

5 (A) to 5 (C) show the captured image 20 and the optical arrangement when the position of the concave defect 120 on the camera photographing surface is near the position where the left edge portion of the light band of the lighting device 1 is reflected. Has been done.

In the measurement systems of FIGS. 5A to 5C, the light emitted from the illumination position 1c on the left side is reflected on the right slope of the concave defect 120 and incident on the camera 2 as compared with the case where the concave defect 120 does not exist. do. On the other hand, as compared with the case where the concave defect 120 does not exist, the light emitted from the illumination position 1d on the right side is reflected on the left slope of the concave defect 120 and then incident on the camera 2.

As a result, as shown in the left figures of FIGS. 5A to 5C, the left slope of the concave defect 120 is brightly reflected as a bright point 32, and the right slope is darkly reflected as a dark point 31.

When there is a concave defect 120 at the position of FIG. 5A, the right slope of the concave defect 120 has the same amount of light as the dark zone portion 22 around the defect on the shooting surface of the camera 2, so that it is a dark spot. 31 is hidden behind the dark zone 22 and cannot be seen. On the other hand, since the left slope of the concave defect 120 is brighter on the shooting surface of the camera 2 than the dark zone portion 22 around the defect, the bright point 32 appears to shine brightly.

When there is a concave defect 120 at the position of FIG. 5B, the right slope of the concave defect 120 comes to the position where the bright band portion 21 is reflected, so that the dark spot 31 appears darker than the periphery of the defect. On the other hand, the light emitted from the brightest part in the central part of the illumination is reflected on the left slope of the concave defect 120, while the image of the left end portion of the light band is reflected around the defect on the photographing surface, and the central part of the optical band. It is a little darker than. Therefore, on the photographing surface, the left slope of the concave defect 120 appears to shine brighter than the periphery as a bright point 32.

When there is a concave defect 120 at the position shown in FIG. 5C, the left slope of the concave defect 120 has the same amount of light as the bright band portion 21 around the defect on the shooting surface of the camera 2, so that the bright point 32 is It is hidden behind the bright band 21 and cannot be seen. On the other hand, since the right slope of the concave defect 120 is darker on the shooting surface of the camera 2 than the portion where the surrounding light band is reflected, the dark spot 31 appears dark.

6 (D) to 6 (F) show the captured image 20 and the optical arrangement when the position of the concave defect 120 on the camera photographing surface is near the position where the right edge portion of the light band of the lighting device 1 is reflected. Has been done.

Contrary to FIGS. 5A to 5C, the right slope of the concave defect 120 appears bright and the left slope appears dark. The appearance on the camera shooting surface is the same. In FIG. 6D, the bright point 32 corresponding to the right slope is hidden by the surrounding bright band portion 21, and the dark point 31 corresponding to the left slope is compared with the surroundings. It looks dark.

In FIG. 6E, the bright point 32 corresponding to the right slope of the concave defect 120 appears brighter than the surroundings, and the dark point 31 corresponding to the left slope appears dark.

In FIG. 4 (F), the dark point 31 corresponding to the left slope of the convex defect 110 is hidden by the surrounding dark zone portion 23, and the bright point 32 corresponding to the right slope appears bright.
<Method of distinguishing between convex and concave defects>
If the images can be taken at the timing when the defects 110 and 120 appear at the positions of FIGS. 3 (B), 4 (E) or 5 (B) and 6 (E), the bright spot 32 or the bright spot 32 due to the right slope of the defects 110 and 120 or The dark point 31 and the dark point 31 or the bright point 32 due to the left slope can be measured at the same time, and it is possible to determine whether the defect is a convex defect or a concave defect based on the relative positional relationship.

However, the location where the defects 110 and 120 actually exist is not determined, and when the measurement is performed at regular intervals, the measurement is performed at the timing when both the dark point 31 and the bright point 32 appear on the same image. What you can do is rare.

Most images have dark spots 31 or bright spots 32, as shown in FIGS. 3 (A) (C), 4 (D) (F), 5 (A) (C), and 6 (D) (F). Only one of them is shown.

Therefore, in the present embodiment, a shooting frame number (hereinafter, also simply referred to as a frame number), which is an image number, is applied as an index indicating the relative position of the work 100 with respect to the illumination light, and the relationship between the position of the dark spot 31 and the frame number is determined. By linear approximation, the relative positional relationship between the dark point 31 and the bright point 32 in the image 20 having the same frame number is obtained, and whether it is a convex defect 110 or a concave defect 120 is determined based on this relative positional relationship.

For the convex defect 110, the relationship between each frame number and the X coordinate of the dark point 31 appearing dark on the bright band portion 21 of the image 20 is represented by a black circle, and the frame number and the bright point appearing brightly on the dark band portions 22 and 23 are represented by black circles. When the relationship between the X coordinates of 32 is represented by white circles and these are plotted on the same graph, the arrangement is along the straight line L1 as shown in FIG. The straight line L1 is an approximate straight line calculated from each dark point 31. In the graph of FIG. 7, the horizontal axis is the frame number and the vertical axis is the X coordinate of the defect.

In FIG. 7, since the amount of movement of the defect 110 between the images is larger than the distance between the right slope and the left slope of the convex defect 110, the difference is difficult to understand.

It is possible to obtain an approximate value in advance for how many pixels the convex defect 110 moves on the photographing surface for each image.

Therefore, as shown in FIGS. 8A to 8F, the images 20 of FIGS. 3A to 3C and 4D to 4F are shifted by an approximate amount of movement obtained in advance. However, the images 40a to 40f cut out with a rectangular frame are used.

In the cropped images 40a to 40f, the X coordinate of the dark point 31 appearing on the bright band portion 21 and the X coordinate of the bright point 32 appearing on the dark band portions 22 and 23 are obtained and plotted against the shooting frame number, respectively. The graph is shown in FIG. When the X coordinate of the dark point 31 is represented by a black circle and the X coordinate of the bright point 32 is represented by a white circle and these are plotted on the same graph, the arrangement is along the approximate straight line L2 calculated from the dark point 31. In the graph of FIG. 9, the horizontal axis is the frame number and the vertical axis is the X coordinate of the defect. Compared with the graph of FIG. 7, the graph of FIG. 9 makes it easier to see the relative positional relationship between the bright point 32 and the dark point 31, and will be described later in this graph.

When the convex defect 110 is reflected in the dark zone portions 22 and 23, the dark spot 31 is hidden in the surroundings and cannot be detected, but the relationship between the position of the dark spot 31 reflected in the bright zone portion 21 and the frame number. By approximating with a straight line L2, the position of the dark point 31 in the dark zone portions 22 and 23 can be estimated.

When the convex defect 110 is present near the left edge of the bright band 21 on the image 20, as in the first half of shooting (first half of the measurement) including the boundary between the bright band 21 and the dark band 22 on the left side, it is bright. The X coordinate of the point 32 is larger than the X coordinate of the dark point 31 in the image 20 having the same frame number. On the other hand, when the convex defect 110 is near the right edge of the bright band 21 on the image 20, as in the latter half of shooting (second half of the measurement) including the boundary between the bright band 21 and the dark band 23 on the right side, the convex defect 110 is located near the right edge of the bright band 21 on the image 20. The X coordinate of the bright point 32 is smaller than the X coordinate of the dark point 31 in the image 20 having the same frame number.

That is, the X coordinate of the bright point 32 for the convex defect 120 is larger than the X coordinate of the dark point 31 of the same frame number obtained from the approximate straight line L2 in the first half of shooting, and the image 20 of the same frame number obtained from the approximate straight line L2 in the latter half of shooting. It is smaller than the X coordinate of the dark point 31 of.

As in the case of the convex defect 10, the concave defect 120 also has the images of FIGS. 5 (A) to 5 (C) and FIGS. 6 (D) to 6 (F) as shown in FIGS. 10 (A) to 10 (F). From No. 20, the images 41a to 41f cut out with a rectangular frame while shifting by the approximate movement amount obtained in advance will be described.

In the cropped images 41a to 41f, the X coordinate of the dark point 31 reflected on the bright band portion 21 and the X coordinate of the bright point 32 reflected on the dark band portions 22 and 23 are obtained and plotted against the shooting frame number, respectively. The graph is shown in FIG. The straight line L3 is an approximate straight line calculated based on each dark point 31, the horizontal axis of the graph in FIG. 11 is the frame number, and the vertical axis is the X coordinate of the defect.

In the concave defect 120, the X coordinate of the bright point 32 is smaller than the X coordinate of the dark point 31 in the image 20 having the same frame number obtained by the approximate straight line L3 in the first half of shooting (first half of measurement), and the approximate straight line in the second half of shooting (second half of measurement). It is larger than the X coordinate of the dark point 31 in the image 20 having the same frame number obtained by L3.

The amount of change in the X coordinate for each frame of the dark point 31 and the bright point 32 is the remainder after correcting the rough movement amount of the convex defect 110 or the concave defect 120 in the cutout images 40a to 40f and 41a to 41f of FIGS. 8 and 10. It represents the amount of movement of the convex defect 110 or the concave defect 120 for each frame.

In order to remove the influence of the remaining movement amount of the dark point 31 and the bright point 32, an approximate straight line (L2 in FIG. 9 and L3 in FIG. 11) is calculated with respect to the X coordinate of the dark point 31, and the bright point at the same frame number. The relative positional relationship between 32 and the scotoma 31 is examined.

When the convex defect 110 or the concave defect 120 is reflected in the dark band portions 22 and 23, the dark spot 31 of the convex defect 110 or the concave defect 120 is hidden in the surroundings and cannot be detected. However, by linearly approximating the coordinates of the dark point 31 appearing in the bright band portion 21, the coordinates of the dark point 31 in the image in which the convex defect 110 or the concave defect 120 is reflected in the dark band portions 22 and 23 can be obtained.

The position of the dark point 31 in the image in which the bright point 32 is detected can be obtained by calculating the X coordinate in the image 20 having the same frame number on the approximate straight lines L2 and L3.

As a result, the relative positional relationship between the dark point 31 and the bright point 32 observed in the image 20 having the same frame number can be obtained, and it is possible to determine whether the convex defect 110 or the concave defect 12.
<Discrimination index>
A specific index for discriminating between the convex defect 110 and the concave defect 120 (hereinafter, also referred to as an unevenness discriminating index) will be described.

The shooting frame number (in this example, the frame number of the cutout image 40b of FIG. 8B or the frame number of the cutout image 41b of FIG. 10B) in which the edge of the light band of the lighting device 1 is first applied to the defect position is obtained from the image. back. Here, the frame number of the cutout image 40b of FIG. 8B or the cutout image 41b of FIG. 10B is referred to as "NoEdge Start". Similarly, the photographing frame number (in this example, the frame number of the cut-out image 40e of FIG. 8 (E) or the cut-out image 41e of FIG. 10 (E)) to which the edge of the light band is applied is obtained from the image. Here, the frame number of the cut-out image 40e of FIG. 8 (E) or the cut-out image 41e of FIG. 10 (E) is referred to as "NoEdgeEnd".

As shown in FIG. 12, a dark spot 31 appears in the bright band portion 21 in the image 20 having a frame number of approximately NoEdgeStart to NoEdgeEnd. Further, in the images 20 having frame numbers of NoStart to NoEdgeStart and NoEdgeEnd to NoEnd, bright spots 32 appear in the dark bands 22 and 23. The frame number "NoStart" represents the frame number of the evaluation start image, and "NoEnd" represents the frame number of the evaluation end image. If the frame number of the image in the center is "No Center",
NoCenter = (NoEdgeStart + NoEgdeEnd) / 2 ... (1)
Will be.

As an example of the unevenness discrimination index, in the first half of the measurement (No Start to No Center), the number of bright points 32 on the upper side of the approximate straight lines L2 and L3 of the dark point 31 is subtracted from the number of bright points 32 on the lower side. In the latter half of the measurement (No Center to No End), the total of the number of bright points 32 on the lower side of the approximate straight line minus the number of bright points 32 on the upper side can be mentioned.

The X coordinate of the bright point 32 in the image 20 of the i-th frame number is Xwihte (i), and the X coordinate of the dark point 31 in the image 20 of the i-th frame number obtained by the approximate straight lines L2 and L3 is Xblack (i). In the image of the i-th frame number in the first half of the measurement, the exponent representing the positional relationship between the bright point 32 and the dark point 31 is expressed by equation (2) for n (i).

In the above equation (2), it is shown that n (i) is +1 when Xwihte (i)> Xblack (i) and n (i) is -1 when Xwihte (i) <Xblack (i). ..

Further, in the i-th frame in the latter half of the measurement, the exponent representing the positional relationship between the bright point 32 and the dark point 31 is expressed by Eq. (3) for n (i).

The unevenness discrimination index H0 is expressed by the equation (4).

If the discrimination index H0 is positive, it can be discriminated as a convex defect 110, and if it is negative, it can be discriminated as a concave defect 120.

As described above, the image 20 having the dark band portions 22 and 23 on both sides of the bright band portion 21 is used, and the first half of the measurement including the boundary portion between the bright band portion 21 and the one dark band portion 22 or the bright band portion 21. Simple and accurate discrimination is performed by the unevenness discrimination index including the number of times that the dark spot 31 and the bright spot 32 are in a predetermined positional relationship in either one or both of the latter half of the measurement including the boundary portion with the other dark band portion 23. It can be carried out.

Further, instead of simply counting the number of bright points 32 existing above or below the approximate straight lines L2 and L3, an index indicating the certainty of the bright points 32 may be integrated as a weight as a discrimination index.

Specifically, as an index showing the certainty, the discrimination index H1 of the equation (5) obtained by integrating the area S (i) of the detected bright point 32 as a weight, and the brightness difference abs (brightness difference abs) between the bright point 32 and the periphery. The discrimination index H2 of the equation (6), which is integrated with ΔL (i)) as a weight, and the distance d (i) from the center of the cutout images 40a to 40f and 41a to 41f to the center of the bright point 32 are integrated as weights ( 7) The discrimination index H3 in the equation may be used. Note that D in equation (7) is the distance from the center of the image to the farthest pixel. Alternatively, the equation (8) in which these are integrated may be used.

By incorporating an index showing such certainty, it is possible to perform discrimination with higher accuracy.

In the discrimination method of the above embodiment, an example in which the relationship between the X coordinate of the dark point 31 and the frame number is linearly approximated is shown, but conversely, the relationship between the position of the bright point 32 in the first half of the measurement and the frame number is linearly approximated. , The concave defect 120 or the convex defect 110 may be determined from the relative positional relationship between the dark point 31 position in the first half of the measurement and the bright point 32 position in the image 20 having the same frame number obtained from the approximate straight line.

Similarly, the relationship between the position of the bright point 32 in the latter half of the measurement and the frame number is linearly approximated, and the position of the dark point 31 in the latter half of the measurement and the position of the bright point 32 in the image 20 having the same frame number obtained from the approximate straight line. It may be determined whether it is a concave defect 120 or a convex defect 110 from the relative positional relationship.

When the relationship between the position of the bright point 32 and the frame number is linearly approximated, the index indicating the certainty of the dark point 31 is used as a weight, as in the case of linearly approximating the relationship between the position of the dark point 31 and the frame number. You may add up.

As yet another discrimination method, the relationship between the position of the bright point 32 in the first half of the measurement and the frame number is linearly approximated, the relationship between the position of the dark point 31 in the first half of the measurement and the frame number is linearly approximated, and the two straight lines are used. It may be determined whether it is a concave defect 120 or a convex defect 110 depending on which of the two is on the top.

Similarly, the relationship between the position of the bright point 32 in the latter half of the measurement and the frame number is linearly approximated, the relationship between the position of the dark point 31 in the latter half of the measurement and the frame number is linearly approximated, and which of the two straight lines is on top. It may be determined whether it is a concave defect 120 or a convex defect 110.

The horizontal axis of each graph, that is, an index indicating the position of the work 100 with respect to the illumination light is used as the frame number, but the position of the work 100 obtained by the rotation speed of the encoder that detects the movement of the belt conveyor on which the work 100 is mounted is used. It may be the horizontal axis. Alternatively, the measurement time may be the horizontal axis.

Further, when the work 100 having the convex defect 110 moves from the left to the right with respect to the camera 2 as shown by the arrow A in FIG. 1, as described above, FIG. 3 (A) → (B). Images are obtained in the order of → (C) → FIG. 4 (D) → (E) → (F). On the other hand, when the work 100 moves in the opposite direction, that is, from the right side to the left side of FIG. 1, the order is FIG. 4 (F) → (E) → (D) → FIG. 3 (C) → (B) → (A). You can get the image with.

The work 100 moves in the forward direction regarding the relationship between the shooting frame number and the X coordinate of the dark point 31 in the bright band portion 21, and the relationship between the frame number and the X coordinate of the bright point 32 in the dark band portions 22 and 23. In this case, the relationship is as shown in FIG. 9, but when the work 100 moves in the opposite direction, the relationship is as shown in FIG. That is, the relative positional relationship between the bright spot 32 and the dark spot 31 in the image 20 having the same frame number is reversed. The straight line L5 is an approximate straight line of the dark point 31 in the bright band portion 21.

Similarly, when the work 100 having the concave defect 120 moves in the forward direction, the order is FIG. 5 (A) → (B) → (C) → FIG. 6 (D) → (E) → (F). An image is obtained. On the other hand, when the work 100 moves in the opposite direction, an image is obtained in the order of FIG. 6 (F) → (E) → (D) → FIG. 5 (C) → (B) → (A). In this case, the work 100 is in the forward direction regarding the relationship between the frame number and the X coordinate of the dark point 31 in the bright band portion 21 and the relationship between the frame number and the X coordinate of the bright point 32 in the dark band portions 22 and 23. When moving to, the relationship is as shown in FIG. 11, but when the work 100 moves in the opposite direction, the relationship is as shown in FIG. Similarly, in the case of the concave defect 120, the relative positional relationship between the bright point 32 and the dark point 31 in the image having the same frame number is reversed. The straight line L6 is an approximate straight line of the dark point 31 in the bright band portion 21.

Therefore, it is preferable to calibrate the relationship between the moving directions of the camera 2 and the work 100 in advance and invert the codes of the discrimination indexes H0 to H3 depending on whether the moving direction is the forward direction or the reverse direction. With such a configuration, even if the moving direction of the work 100 with respect to the camera 2 changes, accurate discrimination can be performed by changing the sign of the unevenness discrimination index.

Further, when the camera 2 is attached by being rotated by θ around its own optical axis, the image is rotated by −θ. In that case, the coordinates (X, Y) of the dark point 31 in the bright band portion 21 or the bright point 32 in the dark zone portions 22 and 23 are obtained, and the coordinates (X) rotated by θ by the following equation (9) are obtained. ', Y') is calculated. The rotated coordinate X'is applied as the X coordinate, and it is determined whether the defect is the concave defect 120 or the convex defect 110. The rotation angle θ of the camera 2 is obtained in advance at the time of calibration. After rotating the image by θ, the coordinates of the bright point 32 or the dark point 31 may be obtained.

With such a configuration, even if the camera 2 is rotated with respect to its own optical axis, the coordinates or the image of the dark point 31 and the bright point 32 on the image 20 are rotated according to the rotation angle, so that the accuracy is correct. It is possible to make a distinction.

As described above, in this embodiment, the work 100 can be photographed while being relatively moved with respect to the illumination light illuminating the camera 2 and the work 100, and the relative movement of the work 100 is stopped. In addition to the fact that it is not necessary to take a picture of the work, a plurality of lighting devices for illuminating the work from a plurality of directions are not required, and one lighting device 1 is sufficient, and the configuration is simplified. Moreover, since the work 100 is not measured for reflection by a line sensor, it is possible to determine whether the surface defect is a concave defect or a convex defect regardless of the surface shape of the measurement site.

In the above embodiment, the method of discriminating the unevenness of the surface defect has been described in the example in which the arrangement of the lighting device 1 and the camera 2 is fixed and the work 100 moves relatively.

However, this method can be applied even when the camera 2 and the work 100 are fixed and the striped illumination light of the illumination device 1 moves relatively. In this case, the lighting device 1 itself may be moved, or the lighting device 1 may be configured by a liquid crystal display device or the like, and the striped lighting pattern may be moved on the display surface. Further, both the work 100 and the illumination light may be moved with a speed difference, and in short, at least one of the work 100 and the illumination light may be moved relative to the other.

When the position of the work 100 is fixed and the illumination light moves, the defect position is fixed. Therefore, the relationship between the coordinates of the dark point 31 and the bright point 32 and the frame number is as shown in the graph of FIG. 15 for the convex defect 110. Therefore, the concave defect 120 is as shown in the graph of FIG. The relationship between the coordinates of the dark point 31 appearing in the bright band 21 and the frame number is approximated by straight lines L7 and L8, and the relative positional relationship between the bright point 32 and the dark point 31 in the first half and the second half of the measurement is obtained. It is possible to discriminate unevenness.

In this case, the horizontal axis of the graphs of FIGS. 15 and 16 may be the frame number or the position (coordinates) of the striped illumination light.

Alternatively, the lighting device 1 and the work 100 may be fixed and the camera 2 may be moved. In this case, the horizontal axis of the graphs of FIGS. 15 and 16 may be the frame number or the position (coordinates) of the camera 2.

1 Lighting device 2 Camera 3 Calculation unit 3a CPU
3b ROM
3c RAM
4 Display 20 Image 21 Bright band 22, 23 Dark band 31 Dark point 32 Bright point 40a-40f Cutout image 41a-41f Cutout image 100 Work 110 Convex defect 120 Concave defect

Claims (18)

A lighting device that illuminates the work with illumination light,
Shooting means and
A means of transportation that moves at least one of the work or illumination light or the means of photography relative to the other.
Equipped with
The work illuminated by the illumination light is photographed by the photographing means while the work or the illumination light or at least one of the photographing means is relatively moved relative to the other by the moving means, and further.
Based on a plurality of images obtained by the imaging, including an image having a bright portion corresponding to an irradiated portion of the illumination light or an image having a dark portion corresponding to a non-irradiated portion of the illumination light, the surface of the work is formed on the bright portion. Both the position of the dark spot appearing as a defect, the position of the bright spot appearing as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means, and either the bright spot or the dark spot. A calculation method for calculating an approximate expression that expresses the relationship with the position,
From the approximate expression calculated by the calculation means, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and the discrimination means for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A device for determining the defect unevenness of a work, which is characterized by being equipped with. The defect unevenness discrimination device for a work according to claim 1, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting. The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point are in a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. Item 2. The device for determining defect unevenness of the work according to Item 1 or 2. The defect unevenness discrimination apparatus for a work according to any one of claims 1 to 3, wherein the sign of the unevenness discrimination index is changed according to the relative moving direction of the work with respect to the photographing means or the illumination light. The defect unevenness determination device for a work according to any one of claims 1 to 4, wherein the coordinates of the dark point and the bright point on the image or the image is rotated according to the rotation angle of the photographing means with respect to the optical axis of the photographing means. As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, the coordinates of the illumination light when the illumination light is moved, and the photographing means. The defect unevenness determination device for a work according to any one of claims 1 to 5, which includes at least one of the coordinates of the photographing means in the case of moving the image. Taking a picture of the work illuminated by the illumination light while moving at least one of the work or the illumination light for illuminating the work or the photographing means relative to the other by the moving means. Steps and
Based on a plurality of images obtained by the photographing step, including an image having a bright portion corresponding to an irradiated portion of the illumination light or an image having a dark portion corresponding to a non-irradiated portion of the illumination light, the work is formed in the bright portion. Both the position of the dark spot appearing as a surface defect, the position of the bright spot appearing as a surface defect of the work in the dark part, and the index indicating the relative position of the work with respect to the illumination light or the photographing means, and either the bright spot or the dark spot. A calculation step to calculate an approximate expression that expresses the relationship with the position of
From the approximate expression calculated by the calculation step, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and a discrimination step for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A method for discriminating defects and unevenness of a work, which is characterized by being provided with. The defect unevenness discrimination method for a work according to claim 7, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting. The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point have a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. Item 7. The method for determining defect unevenness of a work according to Item 7 or 8. The method for determining defect unevenness of a work according to any one of claims 7 to 9, wherein the sign of the unevenness discrimination index is changed according to the relative moving direction of the work with respect to the photographing means or the illumination light. The method for determining defect unevenness of a work according to any one of claims 7 to 10, wherein the coordinates of the dark point and the bright point on the image or the image is rotated according to the rotation angle of the photographing means with respect to the optical axis of the photographing means. As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, the coordinates of the illumination light when the illumination light is moved, and the photographing means. The method for determining defect unevenness of a work according to any one of claims 7 to 11, which includes at least one of the coordinates of the photographing means when moving the light. By moving, the work or the work illuminated by the illumination light is photographed by the photographing means while moving at least one of the illumination light or the photographing means for illuminating the work relative to the other. Based on a plurality of images including an image having a bright portion corresponding to an illuminated portion of illumination light or an image having a dark portion corresponding to a non-irradiated portion of illumination light, the bright portion is used as a surface defect of a work. The position of the dark spot that appears, the position of the bright spot that appears as a surface defect of the work in the dark part, and the position of both the index indicating the relative position of the work with respect to the illumination light or the photographing means and the bright spot or the dark spot. And the calculation step to calculate the approximate expression that expresses the relationship between
From the approximate expression calculated by the calculation step, a concave-convex discrimination index indicating the relative positional relationship between the dark point and the bright point is calculated, and a discrimination step for discriminating the unevenness of the surface defect based on the unevenness discrimination index.
A program that lets your computer run. The program according to claim 13, wherein the unevenness discrimination index includes the number of times that a dark point and a bright point have a predetermined positional relationship in either one or both of the first half of shooting and the second half of shooting. The unevenness discrimination index includes an index indicating the certainty of the dark point or the bright point when the dark point and the bright point are in a predetermined positional relationship in either one or both of the first half of the shooting and the second half of the shooting. Item 13. The program according to item 13. The program according to any one of claims 13 to 15, wherein the sign of the unevenness discrimination index is changed according to the relative moving direction of the work with respect to the photographing means or the illumination light. The program according to any one of claims 13 to 16, wherein the coordinates of the dark point and the bright point on the image or the image is rotated according to the rotation angle of the photographing means with respect to the optical axis of the photographing means. As an index indicating the relative position of the work with respect to the illumination light or the photographing means, the photographing frame number, the rotation speed of the rotary encoder provided in the moving means, the measurement time, the coordinates of the illumination light when the illumination light is moved, and the photographing means. The program according to any one of claims 13 to 17, which includes at least one of the coordinates of the photographing means when moving the light.

Images