JP3159063B2 - Surface defect inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to an apparatus for inspecting a surface defect of an object to be inspected, for example, a painted surface of an automobile body.

[0002]

2. Description of the Related Art A conventional surface defect inspection apparatus is disclosed in, for example, Japanese Patent Application Laid-Open No. 2-73139. These images show a predetermined light and dark stripe pattern on the surface to be inspected, and when there is a defect such as unevenness on the surface to be inspected, the received light image has a difference in brightness (luminance) or a change in brightness (luminance) due to the defect. Is differentiated to detect a defect on the surface of the surface to be inspected.

[0003]

However, the conventional surface defect inspection apparatus as described above has the following problems. That is, in the painting of an automobile body, what is usually called a surface defect is a projection on the painted surface resulting from the application of dust and the like, for example, having a diameter of about 0.5 mm to 2 mm and a thickness of about 0.5 mm to 2 mm. Is about several tens of μm. Since the height (thickness) of the convex portion of this degree is relatively large despite its small diameter, the irregular reflection angle of light increases,
Easy to see. On the other hand, there are irregularities that do not become defects. That is, since the concentration of the paint is not strictly constant due to the eddy convection generated in the process of evaporating the paint solvent, extremely thin (low) irregularities are periodically generated in the thickness of the coating film. The unevenness is, for example, when the interval between the peaks is 1
The height of the unevenness is about several μm. Such extremely thin irregularities are of such a degree that they are usually not noticed and do not become defects. However, it may look like "yuzu" or "orange" depending on the degree of light, so it is called "yuzu skin" or "orange skin". In the conventional example as described above, in order to emphasize and detect a change in luminance, extremely thin unevenness that does not become a defect like the above-mentioned “yuzu skin” is also detected, and this may be erroneously determined as a defect. As described above, in the related art, there is a problem that an extremely thin unevenness that does not become a defect such as so-called “yuzu skin” may be erroneously detected as a defect. The present invention has been made in order to solve the problems of the prior art as described above, and it is possible to detect a more minute defect without erroneously detecting an extremely thin unevenness which does not become a defect such as so-called “yuzu skin” as a defect. It is an object of the present invention to provide a surface defect inspection device that can be performed.

[0004]

Means for Solving the Problems In order to achieve the above object, the present invention is configured as described in the claims. FIG. 1 corresponds to claim 1 of the present invention. In FIG. 1, reference numeral 100 denotes a surface to be inspected;
For example, a painted surface. Reference numeral 101 denotes illumination means for projecting a predetermined light and dark pattern on the surface to be inspected. Also, 102
Is imaging means for imaging the inspected surface 101 and converting the light and dark pattern into image data of an electric signal.
It is a video camera such as a D camera. Reference numeral 103 denotes an image processing unit that extracts only a component having a high frequency component and a level equal to or higher than a predetermined value among the spatial frequency components in the image data obtained by the imaging unit 102 as a defect candidate region. Reference numeral 104 denotes an expansion / contraction unit that performs expansion processing only on the defect candidate area, or performs expansion and contraction processing in a predetermined procedure. Reference numeral 105 denotes a defect determination unit that calculates the area of the defect candidate area after the expansion / contraction processing, and determines whether the defect candidate area is a defect based on the calculated area. These parts 103, 104 and 105 are constituted by a computer, for example. Embodiment 1 which is an embodiment of the present invention is shown in FIGS. Further, in the invention according to claim 2, the expansion / contraction means 104 performs expansion processing only on a defect candidate area having an area satisfying a predetermined condition, or processes expansion and contraction in the order of remaining defect areas. The candidate area is expanded or processed in the order of expansion and contraction. The above embodiment of the invention corresponds to a second embodiment described later with reference to FIGS. Further, in the invention described in claim 3, the expansion / contraction means 104 performs the expansion processing on the defect candidate area in a predetermined direction, or processes the defect candidate area in the order of expansion and contraction. The above embodiment of the invention corresponds to a third embodiment described later with reference to FIGS. Further, according to the invention described in claim 4, while moving one of the inspection surface 100 or the imaging unit 102 and the illumination unit 101,
The image processing unit 103, the expansion / contraction unit 104, and the defect determination unit 105 perform the above-described predetermined processing to detect a defect.

[0005]

As described above, in the invention according to the first aspect, a predetermined light-dark pattern is projected on the surface to be inspected by the illuminating means, which is imaged by the imaging means, and the light-dark pattern is converted into an electric signal image pattern. Convert. Next, the image processing means extracts only a component having a high frequency component and a level equal to or higher than a predetermined value among the frequency components in the image data of the light and dark pattern as a defect candidate. The high-frequency component in the image data is a portion having a change in luminance such as an uneven defect. By extracting only a component having a luminance level equal to or higher than a predetermined value, a candidate region considered as a defect is extracted. Can be. Next, the expansion / contraction means can integrate the light-dark pattern boundary area with the isolated point due to the orange peel generated near the boundary of the light-dark pattern. Next, the defect judging means can distinguish the genuine defect from the light-dark pattern boundary region and detect only the defect. Next, in the second aspect of the present invention, similarly to the first aspect, the expansion / contraction means can integrate the isolated point due to the citron skin generated near the boundary between the light and dark patterns and the light and dark pattern boundary area. Next, in the third aspect of the present invention, similar to the first and second aspects, the expansion / contraction means integrates the isolated point due to the yuzu skin generated near the boundary between the light and dark patterns and the light and dark pattern boundary area. it can. Next, in the invention according to the fourth aspect, the defect is reflected at a position away from the boundary line without being hidden by the boundary in the light and dark pattern, so that the defect can be prevented from being overlooked. In order to inspect the entire surface to be inspected in the invention according to claim 4, the object to be inspected or the illumination means and the imaging means are sequentially moved so that the field of view of the camera scans over the entire surface to be inspected.

[0006]

Hereinafter, the present invention will be described with reference to the drawings. 2 to 5 show the first embodiment. In FIG. 2, reference numeral 1 denotes an illumination device for projecting a predetermined light-dark pattern on the surface 6 to be inspected. Reference numeral 2 denotes a video camera which images the surface 6 to be inspected and converts the light / dark pattern into image data of an electric signal. FIG. 2 shows that there is a defect 7 in the field of view of the camera 2. Reference numeral 3 denotes an image processing device that processes image data obtained by the video camera 2. Reference numeral 4 denotes a computer that performs predetermined processing and expansion / contraction processing on a defect candidate area obtained by processing by the image processing apparatus 3. Reference numeral 5 denotes a computer that performs a defect detection process based on the result of the computer 4.

Next, an example of a defect detection procedure will be described with reference to FIGS. As shown in FIG. 2, when the illumination device 1 irradiates the inspection target surface 6 with a stripe pattern and the reflected light is captured by the monochrome video camera 2, a grayscale image (original image) as shown in FIG. Can be Irregular defect 7
In this case, light is irregularly reflected, so that the defect 7 appears as an area having a different brightness (brightness) from the surroundings as shown in FIG. First, the image processing device 3 inputs an original image (Step S
1). Here, the horizontal direction of the image is x, and the vertical direction is y. In the next step S2, an edge detection process such as differentiation is performed on the original image to extract a region having a luminance change. In step S3, the edge detection image obtained in step S2 is binarized using a threshold value of a predetermined luminance level.
As shown in (b), a binary image is obtained in which the area where the luminance changes is white and the other areas are black. Therefore, a region where the luminance has changed due to the defect 7 and the boundary line between the light and dark patterns and the orange peel is extracted as a white region. At this point, it is not known which white region is defective, so this region is set as a defect candidate region. Subsequently, in step S4, labeling (labeling) is performed on a white region of the image, that is, a defect candidate region. Next, in step S5, the computer 4 calculates the area S of each of the defect candidate areas, and stores the area S for the label of the defect candidate area. Subsequently, a comparison is made with respect to all the labels, that is, the defect candidate areas based on a predetermined area Sref (step S20). If S ≧ Sref, processing is performed in step S6 in the order of expansion and contraction. As shown in the figure, the area of the light-dark pattern boundary line crossing the image y direction is clearly larger than the defect 7 or the citron skin. Therefore,
If the area of the defect to be detected and the area of the light and dark pattern are measured in advance, and an intermediate value is Sref, S
Since only the light / dark pattern boundary line satisfying ≧ Sref is subjected to expansion / contraction processing, the citron skin generated near the light / dark pattern boundary line is integrated (FIG. 3 (c)). Here, the reason why the contraction processing is performed after the expansion processing is to separate the defect again by the contraction processing when the defect is near the boundary between the light and dark patterns and the defect is integrated with the boundary between the light and dark patterns by the expansion processing. Therefore, this shrinking process is not an essential process of the present invention for integrating the yuzu skin and the light and dark pattern boundary line, and may be omitted.

Here, the expansion / contraction processing will be described.
As shown in FIG. 5, when a certain pixel e is set as a target pixel and at least one white pixel image exists among its eight neighboring pixels a to d and f to i as shown in FIG. It is to be. When such processing is performed while scanning the entire image, the white area in the image expands one pixel at a time around the white area. Contrary to the above expansion processing, in contrast to the above expansion processing, when there is at least one black pixel in the eight neighboring pixels, the luminance value of the pixel of interest e is set to black. Will do. The expansion / contraction processing is a general processing for processing the vicinity of 8, but the size of the expansion / contraction processing, that is, how to take the vicinity, etc., is changed according to the state of the inspection surface 6 and the defect 7. There is no problem at all.

Next, the computer 5 calculates again the area of each defect candidate area after the expansion / contraction processing. As a result, a region having a large area becomes a defect in a light and dark pattern boundary line in which citron skin is integrated, and an isolated point having a small area becomes a defect. Is performed, the defect 7 can be detected (step S7) (FIG. 3 (d)). As another embodiment of the present invention, the expansion / contraction processing (step S
Even if the condition in 6) is set to S ≦ Sref, the same effect as in the above embodiment can be obtained because the small-area citron skin region expands and integrates with the light-dark pattern boundary line. Note that the processing procedure, the determination method, and the like are not limited to the present embodiment.

Next, a second embodiment of the present invention will be described with reference to FIGS. In the second embodiment, a defect candidate area satisfying a predetermined condition is expanded /
The contraction processing is performed, and then the remaining defect candidate areas are processed. In the second embodiment, as shown in FIG. 6, first, a region having a large area satisfying S ≧ Sref is expanded / contracted on the basis of Sref (step S6), and then Sr
An area smaller than ef is processed (step S7). Here, if the expansion processing is performed on all the regions at the same time, the defect 7 and the boundary pattern of the light and dark patterns are more easily integrated. However, as in the present embodiment, the expansion / contraction processing is divided into steps S6 and S7. As a result, only the citron skin is integrated with the light-dark pattern boundary line and the area is determined (step S
8) Only the defect can be more reliably detected.
Further, even if a part of the white area of the defect 7 is missing at the time of the binary image (step S3) as shown in FIG.
The defect 7 has a good shape without any missing parts as in the image shown in (d) (the process of step S7). Note that the above processing procedure and the like are not limited to the present embodiment.

Next, a third embodiment of the present invention will be described with reference to FIGS. In the third embodiment, a defect candidate area is expanded / contracted in a predetermined direction. As shown in FIG. 8A, the light-dark pattern boundary line where detection is unnecessary and in which yuzu skin easily occurs crosses in the image y direction. Therefore, if the expansion processing is performed in the image y direction, only the yuzu skin can be integrated with the light-dark pattern boundary line while the defect remains separated (FIG. 8).
(B)). For example, as shown in FIG. 9, this can be realized by performing expansion / contraction processing on four neighborhoods in the image y direction with the pixel of interest being c. Note that the above processing procedure and the like are not limited to the present embodiment.

Next, a fourth embodiment which is an embodiment of the present invention will be described. In the fourth embodiment,
In order to prevent oversight when the defect 7 is on the boundary of the light and dark pattern, the above-described series of defect detection processing is continuously performed while moving either the inspected surface 6 or the video camera 2 or the illumination device 1; The defect 7 is detected. For example, in FIG. 2, the illumination device 1 and the video camera 2 are fixed, and the object to be inspected is moved (not shown) using a conveyor or the like. If the defect detection processing time is sufficiently faster than the moving speed, a plurality of the defect detection processing can be executed while the defect 7 passes through the field of view of the video camera 2. Therefore, even if the defect 7 is hidden by the light-dark pattern boundary in the image at a certain time, the defect 7
Is detected in an image at another time passing through the visual field of the video camera 2 at least once, the defect 7 can be detected. Here, the greater the number of light-dark pattern boundaries that appear in one screen, the higher the probability that the defect 7 will be hidden, so the processing time needs to be shortened. Therefore, the size of the visual field of the video camera 2, the moving speed, and the like may be determined according to the number of light-dark pattern boundary lines reflected on one screen and the processing capabilities of the image processing device 3 and the computers 4 and 5.

As another embodiment, the illumination device 1 and the video camera 2 are sequentially moved by using a robot (not shown) or the like so that the field of view of the video camera 2 scans the entire surface 6 to be inspected. Is also good.

[0014]

As described above, according to the present invention, even if extremely thin irregularities which do not cause a defect such as “yuzu skin” are formed on a non-planar surface to be inspected, only the defect is accurately detected. The effect that it can be obtained is obtained.

[Brief description of the drawings]

FIG. 1 is a functional block diagram of the present invention.

FIG. 2 is an overall diagram showing the first embodiment of the present invention.

FIG. 3 is a diagram showing image processing according to the first embodiment.

FIG. 4 is a flowchart illustrating a control flow according to the first embodiment.

FIG. 5 is an explanatory diagram of the first embodiment.

FIG. 6 is a flowchart illustrating a control flow according to the second embodiment.

FIG. 7 is a diagram illustrating image processing according to the second embodiment;

FIG. 8 is a diagram illustrating image processing according to the third embodiment.

FIG. 9 is an explanatory diagram of the third embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Illumination device 2 Video camera 3 Image processing device 4 Calculator 5 Calculator 6 Inspection surface 7 Defect 100 Inspection surface 101 Illumination means 102 Imaging means 103 Image processing means 104 Expansion / contraction means 105 Defect judgment means

──────────────────────────────────────────────────続 き Continued on the front page (58) Fields surveyed (Int. Cl. ⁷ , DB name) G01B 11/00-11/30 102 G01N 21/84-21/958 G06T 7/00

Claims (4)

(57) [Claims] 1. A surface defect inspection apparatus that irradiates a surface to be inspected with light, creates a light reception image based on light reflected from the surface to be inspected, and detects a defect on the surface to be inspected based on the light reception image. An illumination unit for forming a predetermined light and dark pattern on the surface of the inspection object; an imaging unit for converting a light-receiving image obtained by imaging the inspection surface into image data of an electric signal; and a frequency component in the image data. Image processing means for extracting only a component having a level equal to or higher than a predetermined value in a high frequency region as a defect candidate region, and calculating an area of each of the defect candidate regions to obtain a defect candidate region having an area satisfying a predetermined condition. Expansion / contraction processing means for performing expansion processing only or only expansion / contraction processing in the order, and calculating the area of the defect candidate area after the expansion / contraction processing, and calculating the defect candidate area based on the area. Surface defect inspection apparatus comprising: the defect determining unit determines a defect or not, the. 2. The expansion / contraction means calculates the area of each of the defect candidate areas and performs expansion processing only on the defect candidate areas having an area satisfying a predetermined condition, or
2. The surface defect inspection apparatus according to claim 1, wherein after processing in the order of contraction, the remaining defect candidate area is subjected to expansion processing or processing in the order of expansion and contraction. 3. The surface defect inspection apparatus according to claim 1, wherein the expansion / contraction means expands the defect candidate area in a predetermined direction, or processes the defect candidate area in the order of expansion and contraction. 4. The method according to claim 1, wherein the image processing means, the expansion / contraction means, and the defect determination means perform the predetermined processing to detect a defect while moving the surface to be inspected or any one of the imaging means and the illumination means. 4. The surface defect inspection apparatus according to claim 1, wherein: