JPH0827841B2 - Appearance inspection method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial application] The present invention is directed to color defects on the surface of an inspection object, surface irregularities,
The present invention relates to an appearance inspection method for inspecting appearance defects such as contamination of foreign matter, scratches, and cracks by image processing.

[Prior Art] Conventionally, a spatial region including an inspection object is imaged by an image input means such as a television camera, and the obtained image is converted into a line image based on a density difference of the original image, and then the inspection object is inspected. A visual inspection method for extracting visual defects has been considered. In such an appearance inspection method, it is necessary to extract and inspect lines corresponding to defects and foreign substances from the lines obtained in the line image.

Conventionally, at the time of teaching, as shown in FIG. 8, an internal mask M is formed in a region inside the contour line l of the inspection object O, and a line existing in the region inside the internal mask M is regarded as a defect X _1. I was conducting an inspection. Further, in the same manner, a background mask is formed in a region (background) outside the contour line 1 of the inspection object O, and the line existing in the region in the background mask is regarded as the foreign substance X ₂ for the inspection.

[Problems to be Solved by the Invention] In the above-described conventional method, as shown in FIG. 9 (a), the inspection object O ₁ to
When a large number of O ₄ are present, or when the shape of the inspection object O is complicated as shown in FIG. 9B, there is a problem that mask setting work becomes complicated and the processing time becomes long. It was

An object of the present invention is to solve the above-mentioned problems, and it is an object of the present invention to provide a visual inspection method capable of inspecting for defects and foreign matters without performing a work of setting an inspection region by a mask. It is a thing.

[Means for Solving the Problem] In order to achieve the above object, the present invention traces a line corresponding to a contour line of an inspection target from a line image and sequentially obtains the curvature of the contour line, and the obtained curvature is Erase the lines that match the curvature of the preset standard pattern,
In the remaining line image, a region surrounded by a line is set as a defect part candidate, and a defect and a foreign substance are discriminated from each other in the defect part candidate region.

[Operation] According to the above method, lines corresponding to defects and foreign matters can be extracted without forming a mask, so that the process is easier than in the case where a mask is formed. Is effective in the case where there are a large number of objects in the processing area or the shape of the inspection object is complicated, and the curvature obtained by sequentially tracing the line corresponding to the contour line is used as the curvature of the standard pattern. Since the line recognized as the contour line is erased by matching, almost no part of the contour line is left in the line image in which the line corresponding to the contour line is erased, and it is possible to discriminate between the defect and the foreign matter. It is unlikely that a part of the contour line is mistaken as a candidate for a defective portion when performing.

[Embodiment] In the present invention, it is necessary as a pre-processing to convert an original image obtained by an image input means such as a television camera into a line image as in the conventional case. This preprocessing is performed as follows.

First, the original image P ₀ obtained by imaging the spatial area including the inspection object is a grayscale image, and as shown in FIG. 1A, the inspection object O, the defect X ₁ , and the foreign material X ₂ The image includes. Here, the density of each pixel is represented by, for example, 8 bits and is set to 256 gradations. The process of extracting an edge such as a contour line of the inspection object O from the grayscale image is based on the idea that "the edge portion corresponds to a portion where the density change is large". Therefore, it is general to extract the edge by differentiating the density.
The differentiation process is performed by dividing the grayscale image into 3 × 3 pixel local parallel windows W as shown in FIG. That is,
Pixel E of interest and eight pixels A to D, F around the pixel E
To I to form a local parallel window W, and a vertical density change ΔV of the density of the pixels A to I in the local parallel window W.
And the density change ΔH in the lateral direction are calculated by the following formula, and ΔV = (A + B + C) − (G + H + I) ΔH = (A + D + G) − (C + F + I) Further, the differential absolute value | e _E | and the differential direction value ∠e _E It is calculated by the following formula.

However, A to I indicate the densities of the corresponding pixels. By performing the above calculation for all the pixels of the original image P ₀ , the contour of the inspection object O, the defect X ₁ , the foreign substance X _{2 and the} like where the density change is large and the direction of the change is large. Can be extracted.

Next, thinning processing is performed. The thinning process is performed by paying attention to the fact that the larger the differential absolute value, the larger the density change. That is, an edge having a width of one pixel is extracted by comparing the differential absolute value of each pixel with the differential absolute value of surrounding pixels and connecting those that are larger than the surrounding pixels. . That is, if the position of each pixel on the screen is represented by XY coordinates and the differential absolute value is taken on the Z axis, a curved surface representing the differential absolute value is formed. This is equivalent to finding the ridgeline at. At this stage, edges such as noise are also included, so set the threshold appropriately,
The noise component is removed by using only the value that is equal to or greater than the threshold value.

The line image obtained by the thinning process tends to be a discontinuous line when the contrast of the original image is insufficient or when there is a lot of noise. Therefore, edge extension processing is performed. The edge extension processing starts from the end point of the discontinuity line, compares the pixel of interest with the surrounding pixels, extends the edge in the direction in which the evaluation function f (e _J ) represented by Keep doing this until you hit the end of the line.

Here, e ₀ is the differential data of the central pixel, e _J is the differential data of the adjacent pixel, and J = 1, 2, ...

With the above processing, as shown in Fig. 1 (b), the inspection object O, defect X _1, line images P ₁ to outline l ₀ to l ₂ becomes a pattern of a closed curve of the foreign substance X ₂ and the like to give Be done. next,
A contour inspection for inspecting the presence or absence of a chip, a protrusion, an irregular shape, etc. on the contour line l ₀ of the inspection object O based on the line image P ₁ , and a scratch, a stain on the inside of the contour line l ₀ of the inspection object O, A defect inspection for inspecting the presence or absence of discoloration or the like is sequentially performed.

First, the contour inspection is performed as follows. Here, in the visual inspection, since the inspection objects O of the same kind are inspected one after another, the basic shape and size of the inspection object O are known in advance, and the space area imaged by the image input means is also known. The position in the car is set to the same position every time. Therefore, the contour inspection is performed by collating the pattern of the inspection object O with a preset standard pattern. That is, first, as shown in FIG. 3, search lines a to e that run in a direction intersecting the contour line l ₀ of the inspection object O are set starting from the background portion considered to be in the vicinity of the inspection object O. Then, the intersection between the search lines a to e and the contour line l ₀ is obtained. The intersection of each of the search lines a to e and the contour line l ₀ is
It is obtained according to the procedure shown in FIG. For example, with respect to one search line a, one pixel is advanced from the start point, and if there is a change point of brightness, this point is registered as an intersection with the contour line l _0, and the search line a is advanced to the end point. is there. Similarly, intersections are obtained and registered for each of the search lines b to e. Here, a plurality of (5) search lines a to e are set, the foreign matter X _{1 and the} like exist on the search lines a to e, and a plurality of intersections occur on one search line a to e. Even in such a case, it is possible to identify which intersection is the intersection on the contour line l ₀ . That is, assuming that the foreign matter X ₁ etc. does not exist on all the search lines a to e, only one intersection is present for each intersection on the search lines a to e from which a plurality of intersections are obtained. By examining the direction from the obtained intersections on the search lines a to e, it is possible to identify which intersection is on the contour line l ₀ . In this way, it is possible to avoid the situation where the contour line cannot be specified. Next, the procedure of Figure 5, performs the tracking of the contours l ₀ as a starting point any one of the intersection points on the contour line l _0.
Here is of performing tracking in the direction given by the differentiation direction value for each pixel (indicated by arrows in FIG. 3), sequentially obtains the curvature of the contour line l ₀ while tracing the contour line l _0, The curvature of the corresponding part in the standard pattern is compared, and if the curvature matches, the pixel for which tracking has been completed is deleted. Here, the pixels may be sequentially erased during the tracing of the contour line l ₀ , or may be erased collectively after the tracing is completed. If the curvatures do not match, it is determined that some defect exists on the contour line l ₀ , and the defect is inspected. In this way,
When the contour line l ₀ of the inspection object O is traced without any defect being detected, the line corresponding to the inspection object O is erased from the image as shown in FIG. The pattern of the line image is considered to represent a defect portion corresponding to the defect X ₁ inside the inspection object O and the foreign substance X ₂ outside the inspection object O. That is, at this point, the inspection object O and the defective portion are separated without using a mask as in the conventional case.

By the way, depending on the shape of the edge of the inspection object O and the way of illuminating the inspection object O, an imaginary line l ₀ ′ is shown in FIG.
A line such as shown in may appear, and the contour line may become a double line. However, since the tracing of the contour line l ₀ proceeds in the direction given by the differential direction value, for example, the contour line l ₀ is reversed. If proceeding in the clockwise direction, the contour line l ₀ ′ will occur on the left side of the contour line l ₀ as shown in FIG. Therefore, with respect to the contour line l ₀ for which the intersections with the search lines a to e have been obtained, the contour is provided inside (on the left side) with a width of N pixels corresponding to the width between the double lines l ₀ and l ₀ ′. The lines l ₀ and l ₀ ′ should be erased.

The following processing is performed on the image P ₂ after the contour line l ₀ of the inspection object O is erased as described above with reference to the density of the original image P ₀ . That is, as shown in FIG. 7, the inspection object O is brighter than the background B, the defect X ₁ inside the inspection object O is darker than the inspection object O, and the inspection object O Focusing on the condition that the foreign substance X ₂ existing outside is lighter than the background (the numerical value in FIG. 7 indicates the lightness level), the lightness of the background B and the lightness of the inspection object O are compared. Set the threshold between. Contour line l ₀ of the inspection object O
If the brightness immediately outside the outline of each area is smaller than the above threshold after erasing, the area is judged to be surrounded by the background, so the outline l of the foreign substance X _{2 2} is found, and on the other hand, if the brightness outside the contour of the area is larger than the threshold value, this area is judged to be in the inspection object O, and therefore the defect X ₁ Contour line l ₁
It can be determined that In FIG. 7, for the sake of convenience, the lightness level in each area is made uniform,
In reality, even in one area, the brightness level is usually uneven, so it is desirable to set a representative value for each area.

As described above, it is able also identify the defective X ₁ and the foreign matter X _2, for defects X _1, the concentration distribution and the like of the sum and outline l ₁ near the differential value on the contour line l _1, defect X It is also possible to determine the degree of ₁ .

[Effects of the Invention] As described above, the present invention traces a line corresponding to the contour line of the inspection object from the line image and sequentially obtains the curvature of the contour line, and the obtained curvature of the standard pattern is preset. The line that matches the curvature is erased, the area surrounded by the line in the remaining line image is set as the candidate of the defect portion, and the defect and the foreign matter are discriminated from each other in the area of the defect candidate. Since the lines corresponding to defects and foreign substances can be extracted without forming a mask, the process becomes easier compared to the case where a mask is formed. There is an advantage that it is an effective method when there are a large number or when the shape of the inspection object is complicated. Moreover, since the line recognized as the contour line is erased by matching the curvature obtained by tracing the line corresponding to the contour line with the curvature of the standard pattern, a line image in which the line corresponding to the contour line is erased is obtained. Has an advantage that a part of the contour line is hardly left behind, and there is little possibility that a part of the contour line is erroneously recognized as a candidate of a defective portion when the discrimination determination between the defect and the foreign substance is performed. Further, since a line image in which the contour line of the inspection object is erased is obtained, it is possible to identify other objects on the background such as foreign substances existing around the inspection object in addition to the defect.

[Brief description of drawings]

1 (a) to 1 (c) are explanatory views showing an example of an original image, a line image, and a line image of a defective portion in the present invention, respectively.
FIG. 4 is an explanatory view showing a local parallel window in the same as above, FIG. 3 is an operation explanatory view showing a relationship between the contour line and the search line in the same as above, and FIG. 4 is a procedure for obtaining an intersection of the contour line and the search line in the same as above. FIG. 5 is an operation explanatory view showing the procedure for tracing the contour line in the same as above, FIG. 6 is an enlarged view showing the double-line contour line in the same as above, and FIG. FIG. 8 is an operation explanatory view showing the determination method of FIG.
FIGS. 9 (a) and 9 (b) are diagrams showing an example of an image that was difficult to inspect in the conventional example. l _{0 to} l ₂ ... contour line, O ... inspection object, X ₁ ... defect, X ₂ ... foreign matter.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G06T 7/60 9061-5H G06F 15/70 330 N

Claims (1)

[Claims] 1. An image input device captures an image of a spatial region including an inspection object, and then an original image obtained by the image input device is converted into a line image based on a difference in pixel density. Trace the line corresponding to the contour line of the inspection object and sequentially find the curvature of the contour line, delete the line that matches the curvature of the standard pattern that is set in advance, and delete the remaining line image. An appearance inspection method characterized in that a region surrounded by a line is used as a candidate for a defective portion, and a region that is a candidate for the defective portion is discriminated from a defect and a foreign substance.