JPH0772098A - Visual inspection of welded part - Google Patents

Abstract

PURPOSE:To collectively judge of whether or not a welded part is good on the basis of an image, by detecting the presence/absence of a welded part and detecting the position and a welding strength. CONSTITUTION:Density of each pixel of an image taken by an image input device 1 is A/D converted 2 and processed in a pretreatment part 3. In other words, a differential image (in a differential direction) is formed on the basis of an original image which has a differential value corresponding to a changing ratio (changing direction) of density of a pixel to that of an adjacent pixel as the value of the pixel. Pixels having larger differential values than those of surrounding pixels are coupled to extract edges of a width of one pixel to obtain an edge image. The obtained images are stored in memories 4-7 respectively for the original image, differential image, differential direction image and edge image. An operating/processing part 8 detects, on the basis of the images, the presence/absence of a welded part, the presence/absence of the welded part at a predetermined position of a to-be-inspected object, and a welding strength from the gloss of the part. Accordingly, an inspection for parts other than the welded part is avoided, and whether or not the position and strength of the welded part are good can be judged.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a welded portion appearance inspection method for determining the quality of a welded state by performing image processing on an image of an inspection object which is a welded object.

[0002]

2. Description of the Related Art As a technique for determining the quality of a welded state, the present inventors have obtained a frequency distribution of differential direction values for pixels on a boundary line of a welded place based on an image of an inspection object including the welded place. A method has been proposed in which the quality of the welded state is judged by comparing this frequency distribution with the frequency distribution of a non-defective product which is obtained in advance (Japanese Patent Laid-Open No. 5-54).
No. 107).

[0003]

The above-mentioned technique detects the presence or absence of a welded portion and judges the quality based on the frequency distribution of the differential direction values reflecting the welded shape.
Even if it is not a welded part, there is a possibility that it will be mistaken for a good product if there is a part with the same shape as a good product, and there is a problem that it is impossible to judge the quality of the welding position deviation and welding strength etc. .

The present invention is intended to solve the above-mentioned problems, and to identify a welded portion on the basis of an image to accurately specify a portion to be recognized, and also regarding positional deviation of the welded portion and welding strength. An object of the present invention is to provide a welded part appearance inspection method that enables the quality of a welded part to be comprehensively determined by enabling the quality determination.

[0005]

According to a first aspect of the present invention, welding is performed in a state where light is applied to an inspection object having a welded portion so that a difference in illuminance occurs between the welded portion and a peripheral portion of the welded portion. A process of detecting the presence / absence of a welded part in a welded part appearance inspection method in which a region including a part is imaged and the imaged image is subjected to image processing to evaluate the welded state based on the appearance of the welded part. The method is characterized by including a step of determining whether or not the portion is located at a required position of the inspection object, and a step of determining the welding strength based on the gloss of the welding portion.

According to a second aspect of the present invention, in the first aspect, based on the original image that is a grayscale image, a differential image in which a differential value corresponding to a rate of change in density between adjacent pixels is used as a value of each pixel,
A differential direction image in which the differential direction value corresponding to the change direction of the density with the adjacent pixel is used as the value of each pixel, an inspection window including a plurality of pixels is set in the image, and the pixels in the inspection window are scanned lines. Is moved and scanned to find edge candidate points that are considered to be the edges of the welding site based on at least one of the differential value and the differential direction value of each pixel, and then the adjacent edge candidate points are differentiated and the differential direction. It is characterized in that it is sequentially obtained on the basis of at least one of the values and it is determined that a welding site exists in the inspection window when the number of edge candidate points exceeds a prescribed threshold value.

According to a third aspect of the present invention, in the second aspect, a direction search window is further set in the inspection window,
The pixels in the direction search window are scanned to obtain the frequency distribution of the differential direction value for the pixels whose differential value of each pixel is greater than or equal to the specified threshold, and the scanning line is set along the direction with the highest frequency, It is characterized in that the smaller number of directions is the moving direction of the scanning line.

According to a fourth aspect of the present invention, a terminal reference line serving as an edge of one of the welded members is set, and the total number of pixels between the terminal reference line and the edge candidate point obtained by the method of the second aspect is set. It is characterized in that the suitability of the position of the welded part is determined by obtaining this value and comparing this total number with a prescribed threshold value. According to the invention of claim 5, a terminal reference line which is an edge of one of the welded members is set, and the edge candidate point obtained by the method of claim 2 is continuous between the welded edge line and the terminal reference line. Is determined and the suitability of the position of the welded portion is determined by comparing this area with a prescribed threshold value.

According to a sixth aspect of the present invention, in the fourth or fifth aspect, the edge of one member welded based on at least one of the differential value and the differential direction value of each pixel by scanning in the inspection window. The above pixel is obtained, and when a specified number or more of pixels satisfying the same condition are continuously present, the consecutive pixels are used as the terminal reference line. According to a seventh aspect of the present invention, in the fourth or fifth aspect of the invention, the density of the pixels on the scanned line is scanned in a direction substantially orthogonal to the edge of one of the welded members starting from the prescribed search start point. A threshold value for density is set based on the tendency of increase and decrease of, and a pixel whose density level is inverted with respect to the threshold value on the scanned line is used as a terminal point, and the pixel passes through two terminal points obtained by the same procedure. It is characterized in that a straight line is used as the terminal reference line.

According to an eighth aspect of the present invention, in the welded portion visual inspection method according to the seventh aspect, when there is no terminal point on a line scanned starting from the search start point, a predetermined pixel portion is formed in a direction orthogonal to the scanning direction. It is characterized in that a new shifted start point is set and a terminal point is obtained for this new start point. According to the invention of claim 9, the total number of pixels between each edge candidate point obtained by the method of claim 2 and a prescribed side of the inspection window set at a fixed position with respect to the inspection object is calculated, and this total number is calculated. It is characterized in that the suitability of the position of the welded portion is determined by comparing with a prescribed threshold value.

According to a tenth aspect of the present invention, between the weld edge line connecting the edge candidate points obtained by the method of the second aspect and a prescribed side of the inspection window set at a fixed position with respect to the inspection object. Is determined and the suitability of the position of the welded portion is determined by comparing this area with a prescribed threshold value. The invention of claim 11 relates to claim 1.
In the invention, the received light amount corresponding to each pixel is obtained in the inspection window set to include the welding part in the original image which is a grayscale image, and the specified offset value is added to the average light amount of all the pixels in the inspection window. An average amount of light is calculated for a pixel having a smaller amount of received light than a threshold value, and when the average amount of light is larger than a prescribed threshold value, the gloss is insufficient and the welding strength is determined to be poor.

According to a twelfth aspect of the present invention, in the first aspect of the invention, the amount of received light corresponding to each pixel is obtained within the inspection window set so that the original image, which is a grayscale image, includes the welding portion, It is characterized in that the welding strength is determined to be non-defective when the total number of pixels whose received light amount is larger than the threshold value obtained by adding the specified offset value to the average light amount of all pixels is larger than the specified threshold value.

According to a thirteenth aspect of the present invention, in the first aspect of the present invention, a mask for defining the search direction is set so that the original image, which is a grayscale image, includes a welding portion, and the amount of received light is a threshold value or more. Is determined to be non-defective in terms of welding strength when a predetermined number or more continuously exist in the search direction. According to a fourteenth aspect of the invention, in the invention of the eleventh to thirteenth aspects, the light source for which specular reflection light is imaged and the light source for which diffuse reflection light is imaged are turned on at different timings, and light from each light source is changed. It is characterized in that the quality of the gloss of the welded part is determined based on the above.

According to a fifteenth aspect of the present invention, in the first aspect of the present invention, there are provided a determination result of presence / absence of a welded portion, a determination result of appropriateness of a position of the welded portion, and a determination result of welding strength based on gloss of the welded portion. It is characterized in that the operation of the welding apparatus is corrected and the welding position is corrected using at least one of them.

[0015]

In the present invention, the process of detecting the presence / absence of a welded part, the process of determining whether or not the welded part is located at the required position of the inspection object, and the weld strength based on the gloss of the welded part are determined. By having the process, it is possible to make a comprehensive judgment on the welded part and to make an accurate judgment on the quality of welding.

Further, since it is possible to make an accurate determination regarding welding, it is possible to manage the operation of the welding apparatus and correct the welding position, and to improve the quality of products having welded parts, that is, to improve the yield. It connects.

[0017]

[Example] The welding site to be inspected is irradiated with light from an oblique direction so that a difference in the illuminance between the welding site and the surrounding occurs and a shadow is formed, and the region including the welding site is the area of the ITV camera. An image is captured by such an image input device. Therefore, contrast can be obtained in the image captured from the image input device. The present invention utilizes such image contrast to determine the quality of the welded state.

As shown in FIG. 6, the density of each pixel of the image captured by the image input device 1 is converted into a digital signal in the analog-digital converter 2, and the following preprocessing is performed in the preprocessor 3. By doing so, in addition to the original image obtained as the output of the analog-digital conversion unit 2, a differential image, a differential direction image, and an edge image are obtained.

That is, the original image obtained by imaging the spatial region including the inspection object is a grayscale image, and the processing for extracting the edge such as the contour line of the inspection target from the grayscale image is
It is based on the idea that "the edge corresponds to a portion where the density change is large". Therefore, the edge is extracted by differentiating the density. The differential processing is shown in FIG.
The original image P ₁ is divided into 3 × 3 pixel local parallel windows W as shown in FIG. That is, with the pixel E of interest in the original image P ₁ as the center, the eight pixels (8 neighborhoods) A to D and F to I around the pixel E have a local parallel window W.
And the vertical density change ΔV and the horizontal density change ΔH of the pixels A to I in the local parallel window W are calculated by the following equation.

ΔV = (A + B + C)-(G + H + I) ΔH = (A + D + G)-(C + F + I) where A to I represent the densities of the corresponding pixels. Furthermore, the differential value abs (E) for pixel E and the differential direction value de
g (E) and are calculated by the following equation. abs (E) = (ΔV ² + ΔH ² ) ^1/2 deg (E) = tan ⁻¹ (ΔV / ΔH) + (π / 2) The differential value abs (E) is in the vicinity of the pixel E of interest in the original image. It represents the rate of change of density in the region, and the differential direction value deg (E) is
The direction orthogonal to the direction of density change in the vicinity region is shown. By performing the above calculation for all the pixels of the original image P ₁ , it is possible to extract a portion such as the contour line of the inspection object where the density change is large and the direction of the change. Here, the value of each pixel is the differential value abs (E)
Is called a differential image, and an image having a differential direction value deg (E) is called a differential direction image.

Next, a thinning process is performed. The thinning process is performed by paying attention to the fact that the larger the differential value, the larger the density change. That is, the differential value of each pixel is compared with the differential value of the surrounding pixels, and those having a larger value than the surrounding pixels are connected to extract an edge having a width of one pixel. As shown in FIG. 8, considering a differential image P _{2 in} which the position of each pixel is represented by XY coordinates and the differential value is represented by Z coordinates, the thinning processing corresponds to finding the ridgeline on this curved surface. . By the processing up to this point, all ridgelines are extracted regardless of the magnitude of the differential value. Since the ridge line obtained at this stage also includes unnecessary small mountains due to noise or the like, FIG.
As shown in, the threshold value SL is appropriately set, and only the value equal to or higher than the threshold value SL is adopted to remove the noise component.
The image obtained by this processing 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. In the edge extension processing, starting from the end point of the discontinuous line, the pixel of interest and surrounding pixels are compared, and the line is extended in the direction in which the evaluation function f (J) represented by the following equation becomes the largest,
Continue doing this until you hit the endpoint of the other line.

F (J) = abs (J) .cos (deg (J) -deg (0)).
cos ((J−1) π / 4−deg (O)) where deg (O) is the central pixel (E of the local parallel window W).
Deg (J) is the differential direction value of an adjacent pixel (corresponding to 8 neighborhoods of the local parallel window W), ab
s (J) is the differential value of the adjacent pixel. However, J = 1, 2,
…, 8. Through the above processing, an edge image that traces the contour of the portion of the original image P _{1 where} the density change is large can be obtained. Thus, as shown in FIG. 10, the original image P ₁ ,
Differential image P ₂ , differential direction image P ₃ , and edge image P ₄ 4
Images of different types are obtained, and each image is stored in the original image memory 4, the differential image memory 5, the differential direction image memory 6, and the edge image memory 7, respectively. In the following description, each image P ₁
The pixel positions of ~ P ₄ are represented by XY coordinates, and
The pixel values in each image are f ₁ (x, y) and f, respectively.
Let ₂ (x, y), f ₃ (x, y), and f ₄ (x, y).

The original image is a grayscale image, and the density a (= f ₁ (x, y)) of each pixel is represented by, for example, 8 bits and 0 ≦ a ≦ 255. Further, the differential value b (= f ₂ (x, y)) of each pixel of the differential image is represented by 6 bits, for example, 0 ≦ b ≦ 63, and the differential direction value c (= of each pixel of the differential direction image f ₃ (x, y) is expressed in 16 directions, for example, and 0 ≦ c ≦ 15. As for the edge image, only the presence / absence of a line is present, and therefore the value of the pixel forming the line is “1” and the value of the other pixels is “0”. That is, f ₄ (x, y)
The value range of is {0, 1}.

The arithmetic processing section 8 is a section for performing image processing by a microcomputer, and is a section for performing the processing described below. The processing in the arithmetic processing unit 8 is basically based on the process of detecting the presence / absence of a welding site, the process of determining whether or not the welding site is located at a required position of the inspection object, and the gloss of the welding site. And the process of determining the welding strength. By these three processes, it is possible to prevent inspection of parts other than the welded part, and it is possible to judge the position of the welded part and the weld strength. The processing procedure for each process will be described below.

Determination of Presence / Absence of Welded Site (Method 1-1) To determine the presence / absence of a welded site, an edge image, a differential image, and a differential direction image are used. First, FIG.
Welded portion 12 of inspection object 11 welded as shown in FIG.
A rectangular inspection window W ₁ as shown in FIG. 3 is set in the captured image. The scan line L ₁ is set in the inspection window W ₁ in the horizontal direction, the scan line L ₁ is scanned upward pixel by pixel, and the scan line L ₁ is scanned at each position.
Scan the pixel above ₁ . That is, the inspection window W
Raster scanning will be performed in ₁ . At this time, if the scanning line L ₁ is scanned, a pixel forming an edge in the edge image (that is, a pixel whose value is “1” in the edge image; hereinafter referred to as edge flag) X is scanned line. L
_{Since 1} touches, the differential value and the differential direction value are obtained for this edge flag X. The condition that the differential value is greater than or equal to the specified threshold value, and the differential direction value is within the specified range (here, since the scanning line L ₁ is scanned in the upward direction, the differential direction value is the lateral direction value). If the condition of “get” is satisfied, it is highly possible that the pixel is the edge of the welded part, and this pixel is set as the edge candidate point Y.

Next, as shown in FIG. 4, the rectangular search windows W _L and W _R are set to the left and right so that the obtained edge candidate point Y is a corner pixel. Here, the numbers of pixels of the search windows W _L and W _R are the same in the vertical and horizontal directions in FIG. 4, but the numbers of pixels may be different in the vertical and horizontal directions. When the search windows W _L and W _R are set in this way, the scan line L ₁ is scanned within the range of the search windows W _L and W _R ,
The differential value and the differential direction value of the edge flag X of the edge image are evaluated to obtain the edge candidate point Y. When the scanning of the scan line L _{1 in} the search windows W _L and W _R is completed, in the right search window W _R , the edge candidate point Y at the right end is set as a corner pixel as shown in FIG. The next search window W _R2 is set as follows, and the left search window W _{L is set.}
If so, the next search window is set such that the left edge candidate point Y is a corner pixel.

When the edge candidate points Y indicate the boundary line of the welded portion 12, the edge candidate points Y should be lined up continuously, so that the above-described processing is repeated to obtain continuous edge candidates. The point Y can be obtained. If the edge candidate point Y does not exist in the set search windows W _L and W _R , or if the search windows W _L and W _R go out of the inspection window W ₁ , the search for the edge candidate point Y is stopped. Then, the number A of edge candidate points Y from the time when the first edge candidate point Y is found is obtained.

When the number A is larger than the specified threshold value S (A> S), it is determined that the obtained continuation of the edge candidate points Y represents the boundary line of the welded portion 12. Also, A
When> S is not satisfied, the scanning line L ₁ is continuously scanned for the unscanned portion of the scanning line L _{1 in} the inspection window W ₁ , and the same processing is performed. Here, if the condition of A> S is not satisfied even after the scan line L ₁ is scanned over the entire inspection window W ₁ , the welded portion 12
Is determined not to exist in the inspection window W ₁ .

As described above, the inspection window W ₁
It is determined that the welded portion 12 exists when it is determined whether the edge flag X of the edge image is the edge candidate point Y, and when the number of consecutive edge candidate points Y is equal to or greater than the specified threshold value. Therefore, the presence or absence of the welded portion 12 can be reliably detected. The above operation is summarized as shown in FIG. That is, the scanning line L ₁ is moved pixel by pixel in the set inspection window W ₁ (S1), and the edge flag is searched for on the scanning line L ₁ . An edge flag whose differential value is greater than or equal to a specified threshold value in the edge image and whose differential direction value is within a specified range is obtained (S2), and the edge flag closest to the scanning start position of the scanning line L ₁ is determined. It is extracted as an edge candidate point Y (S3). In the vicinity of this edge flag, the search window W
By setting _R and W _L and performing scanning in each search window W _R and W _L , the differential value in the edge image is greater than or equal to the specified threshold value, and the differential direction value is within the specified range. A certain edge flag is obtained as the edge candidate point Y (S4).
Further, the search windows W _R and W _L are sequentially set and the search for the edge candidate point Y is continued (S5). Setting the search window W _R, or not an edge flag is present in the W _L, discontinue the search for the edge flag when the set search window W _R, where W _L goes out from the inspection window W _1, it The number of edge candidate points Y obtained up to this point is obtained (S6). The obtained number A of edge candidate points Y is compared with a prescribed threshold value S, and if A> S is satisfied, it is determined that the obtained edge candidate points Y are the boundaries of the welded portion 12 (S7). ).

In this method, the edge candidate point Y is determined when both the differential value and the differential direction value are satisfied, but the condition is satisfied only for one of the differential value and the differential direction value. In this case, the edge candidate point Y may be determined. (Method 1-2) This method shows a method of automatically selecting the moving direction of the scanning line L ₁ in step S1 in FIG. In this process, a direction search window W ₂ as shown in FIG. 11 is further set in the inspection window W ₁ . Then, to scan through the direction search window W ₂ is set, after extracting all edges flag in the direction search window W _2, the differential value is defined for the edge flag differential direction value is not less than the threshold value Find the frequency distribution of. Here, as described above, the differential direction value is represented by 16 directions, and if each differential direction value is assigned as shown in FIG. 12, for example, the differential direction value is 0,
In the range of 1, 14, 15, 6, 7, 8, and 9, the direction of the boundary line of the welded portion 12 is considered to be substantially horizontal,
In the range of 2, 3, 4, 5, 10, 11, 12, and 13, it is considered that the direction of the boundary line of the welded portion 12 is substantially vertical. Therefore, a guideline for determining whether the boundary line of the welded portion 12 is the horizontal direction or the vertical direction is obtained from the frequency distribution, and if the horizontal direction, the scanning line L ₁ in the vertical direction is scanned in the horizontal direction, and if it is the vertical direction. The direction of the scanning line L ₁ and the moving direction are automatically set so that the scanning line L _{1 in} the horizontal direction is scanned in the vertical direction. Other techniques are the same as those in Method 1-1.

Judgment of Welding Position (Method 2-1) After judging whether or not it is the welded portion 12,
It is determined whether the obtained position of the welded portion 12 is a predetermined position. In this process, first, the positions of the edge candidate points Y obtained in the process of obtaining the welded portion 12 are stored in the buffers whose regions are set in the memory. Next, FIG.
As shown in FIG. 4, scanning is started from two preset search start points st1 and st2 in the vertical direction or the horizontal direction so as to be substantially orthogonal to the boundary line of the inspection object 11.
In this scanning, an edge flag in an edge image in which the differential value is equal to or larger than a specified threshold value and the differential direction value is within the specified range, and the obtained edge flags are respectively set to the terminal points p1 and p.
Set to 2. Assuming that the edge of the inspection object 11 is a straight line, the straight line passing through the obtained terminal points p1 and p2 is considered to be the edge of the inspection object 11, and thus this straight line is used as the position coordinates of the terminal points p1 and p2. The terminal reference line L ₂ is obtained based on Since the terminal reference line L ₂ must be set so as to pass through the welded portion 12, the search start points st1 and st2 are set so that the terminal reference line L ₂ satisfying the condition can be obtained.

If the position of the welded portion 12 is the desired position,
It is essential that the area welded part 12 occupies one side of the terminal base line L ₂ when the cut welded part 12 in the terminal base line L ₂ is within a predetermined range. Therefore, the edge candidate points Y that are arranged on one side across the terminals reference line L ₂ among the edge candidate point Y of the welded part 12, as shown in FIG. 15, respectively the distance between the terminal reference line L ₂ It is calculated as the number of pixels between each edge candidate point Y and the terminal reference line L _2, and the total of the calculated number of pixels is calculated. This total number of pixels reflects the area of the portion surrounded by the curve connecting the edge candidate points Y and the terminal reference line L ₂ .

Therefore, the total number B of pixels is compared with a prescribed threshold value MS, and if B> MS is established, it is determined that the welding site 12 is a normal position on the inspection object 11. If the above conditions are not satisfied, it can be determined that the position of the welded portion 12 is defective. The above procedure is summarized in FIG. That is, scanning is started from two search start points st1 and st2 to obtain two terminal points p1 and p2 on the edge of the inspection object 11, and the terminal reference line L ₂ is set so as to pass through both terminal points p1 and p2. Set (S1). Next, the number of pixels between the terminal reference line L ₂ and each edge candidate point Y is calculated (S2), and the sum B of this number of pixels is calculated (S3). If the total number B of pixels is larger than the specified threshold value MS, the welded portion 12 is determined to be non-defective (S4).

(Method 2-2) This method is basically Method 2.
The procedure for obtaining the terminal reference line L ₂ is the same as that for Method 2-1. Further, in this method, as shown in FIG. 16, the weld edge line L is formed by connecting the edge candidate points Y.
Ask for ₃ . The weld edge line L ₃ obtained in this way
The area C surrounded by the terminal reference line L ₂ is calculated, and the obtained result is compared with a prescribed threshold value MS in the same manner as in Method 2-1, and when C> MS is satisfied, The position of the welded portion 12 is determined to be normal. That is, the area C is to be obtained more accurately than the method 2-1 in which the area is obtained by the number of pixels, and the other points are the same as the method 2-1.

(Method 2-3) Method 2-1 and Method 2-
2, the terminal reference line L ₂ was obtained as a straight line passing through the two terminal points p1 and p2, but in this method, as shown by the arrow in FIG. 17, scanning is performed in the prescribed direction within the inspection window W _1. Then, an edge flag whose differential value is equal to or larger than a specified threshold value in the edge image and whose differential direction value is within a specified range is obtained. This edge flag is assigned to the terminal points p1 and p
Although it is equivalent to 2, instead of setting a straight line so as to pass through the obtained edge flag, as shown in FIG.
A connectivity confirmation window W ₃ having a predetermined number of pixels (a plurality of pixels in the vertical and horizontal directions) is set so as to include the obtained edge flag Z, and in the connectivity confirmation window W ₃ , the differential value is equal to or more than a specified threshold value. , And the edge flag Z in the edge image whose differential direction value is within the specified range. Thus, the continuous edge flag Z is obtained, and the continuous edge flag Z
When the number D of the above is D> NS with respect to the prescribed threshold NS, the continuous edge flag Z is set as the terminal reference line L ₂ . If the above conditions are not satisfied, by continuing scanning in the inspection window W ₁ ,
The terminal reference line L ₂ is obtained. Other points are Method 2-1
Is the same as.

(Method 2-4) Method 2-1 to Method 2-
In FIG. 3, the edge flag is used to obtain the terminal reference line L ₂ , but the edge flag can be obtained only when the contrast between the inspection object 11 and the background is relatively large. If it is small, the edge flag may not be surely obtained. Therefore, in this method, a procedure for obtaining the terminal reference line L ₂ without using the edge flag will be shown.

That is, as shown in FIG. 20, starting from two preset search start points st1 and st2, a predetermined range is scanned vertically or horizontally so as to be substantially orthogonal to the boundary line of the inspection object 11. . Here, the density (light amount) of each pixel on the line to be scanned is obtained from the original image, and it is determined in advance whether the density is increasing or decreasing. In this way, the threshold value PS for determining the boundary line of the inspection object 11 is set by considering the tendency of increase and decrease of the density. The threshold value PS is set higher than the densities of the search start points st1 and st2 if the density tends to increase, and set lower if the density tends to decrease. Next, as shown in FIG. 21, a pixel in which the level of the density first inverts with respect to the threshold value PS starting from the search start points st1 and st2 on the scanning line is obtained. That is, when the density tends to increase, the pixel whose density exceeds the threshold PS is obtained, and when the density tends to decrease, the pixel whose density becomes smaller than the threshold PS is obtained. In this way, two pixels are obtained and both pixels are regarded as the terminal points p1 and p2 on the boundary line of the inspection object 11.

The subsequent procedure is the same as that of the method 2-1.
A straight line passing through both terminal points p1 and p2 is set as a terminal reference line L ₂ ,
The position of the welded portion 12 with respect to the terminal reference line L ₂ is determined. As described above, in this method, the boundary line of the inspection object 11 is determined only by the density change of the pixel without using the edge flag to obtain the terminal reference line L _{2, so} that the contrast is small and the edge flag is not obtained. It is possible to obtain the terminal points p1 and p2 even in such a place.

The above processing is summarized as shown in FIG. That is, first, scanning is started from the search start points st1 and st2 toward the inspection object 11, the density of each pixel on the scanned line is obtained, and at the same time, the tendency of increase or decrease in density is obtained (S1). Next, if the density tends to increase, a threshold value PS higher than the density at the search start points st1 and st2 is set, and if the density tends to decrease, a threshold value PS lower than the density at the search start points st1 and st2 is set. Set (S2). If the density tends to increase, the first pixel on the line scanned from the search start points st1 and st2 where the density exceeds the threshold value PS is the terminal point p1 or p2. If the density tends to decrease, the density is lower than the threshold value PS. The first pixel that becomes smaller is set to the terminal points p1 and p2 (S3).

(Method 2-5) In Method 2-4, the search start points st1 and st2 are set to two points, but in some cases, the terminal points p1 and p2 cannot be determined for both search start points st1 and st2. There is also. Therefore, when the two terminal points p1 and p2 cannot be determined, the search start point s obtained by shifting the positions of the search start points st1 and st2, which cannot be determined, by a predetermined number of pixels in the direction orthogonal to the scanning direction.
Set t1 and st2 and follow the same procedure to set terminal points p1 and p2.
To ask for. In this way, the two terminal points p
By repeating the above procedure while shifting the search start points st1 and st2 until 1 and p2 are obtained, the terminal reference line L ₂ can be reliably obtained. Other procedures are the same as those in Method 2-4.

(Method 2-6) In this method, the terminal reference line L
Instead of setting ₂ , the prescribed one side of the inspection window W ₁ set at a fixed position with respect to the inspection object 11 is set as a reference line L ₄ and each edge candidate point Y is set as shown in FIG.
For each determined distance between the reference line L ₄ as the number of pixels between each edge candidate point Y and the reference line L _4, obtaining the sum of the number of pixels determined. This sum is the edge candidate points Y at both ends.
From the base line L ₄ to the reference line L ₄ and the edge candidate point Y
It reflects the area of the part surrounded by the curve connecting the lines and the reference line L ₄ . Here, the inspection window W
Depending on how ₁ is set, there may be a plurality of edge candidate points Y on a straight line orthogonal to the reference line L ₄ , but in that case, among the edge candidate points, pixels between the reference line L ₄ and Use the one that maximizes the number.

The total number of pixels B'obtained as described above is compared with a prescribed threshold value MS ', and if B'> MS 'is established, the welded portion 12 is on the inspection object 11 at a normal position. You can judge. If the above conditions are not satisfied, the welded part 12 is determined to be defective. In short, the method is different from Method 2-1 in that one side of the inspection window W ₁ is used as the reference line L ₄ instead of setting the terminal reference line L ₂ , and the other procedure is Method 2-1.
Is the same as.

The above procedure is summarized as shown in FIG. That is, the number of pixels between each edge candidate point Y and the reference line L ₄ is obtained by using one side of the inspection window W ₁ as the reference line L ₄ (S1), and the sum B ′ of the number of pixels is obtained (S2). If the total B'of the number of pixels is larger than the prescribed threshold value MS ', the welded portion 12 is judged to be non-defective (S3).

Terminal points p1 and p2 as in method 2-1
In the case of obtaining, the position of the terminal reference line L ₂ is not constant, so even if the same inspection object 11 is inspected, the result may vary, but the position of the inspection window W ₁ with respect to the inspection object 11 If the appropriateness of the position of the welded portion 12 is determined by this method with constant, the result will be constant, and the position determination of the welded portion 12 can be performed with higher accuracy.

(Method 2-7) In this method, as shown in FIG. 25, one side defined by the inspection window W ₁ set at a fixed position with respect to the inspection object 11 is set as a reference line as in the method 2-6. L ₄ and similarly by connecting edge candidate point Y and ,, method 2-2 obtains the weld edge line L _3, the area of the portion between the weld edge line L ₃ and the reference line L ₄ obtained C ′ Is calculated and compared, and the obtained result is compared with a prescribed threshold value MS ′ in the same manner as in method 2-1, and when C ′> MS ′ is satisfied,
The position of the welded portion 12 is determined to be normal.
Further, if this condition is not established, it is determined to be defective. Other procedures are the same as those in Method 2-1.

The above procedure is summarized as shown in FIG. That is, one side of the inspection window W ₁ is set as the reference line L _4, and the area C ′ (may be the number of pixels) of the portion surrounded by the weld edge line L ₃ connecting the edge candidate points Y and the reference line L ₄ is obtained. (S1) If the area C'is larger than the prescribed threshold value MS ', the welded portion 12 is judged to be non-defective (S2).

Judgment of Welding Strength (Method 3-1) To judge the welding strength, the gloss of the welded portion 12 is judged. That is, when a lot of unevenness is formed in the welded part 12 due to the inclusion of soot, diffuse reflection occurs on the surface when the welded part 12 is illuminated, so that the welded part 12 becomes bright on the image. When the portion 12 is good, the surface of the welded portion 12 becomes a mirror surface and specular reflection occurs, so that the property that part of the welded portion 12 is bright and the other portion is dark is used.

First, the inspection window W ₁ including the welded portion 12 is set, and the average light amount (average density) E of all the pixels in the inspection window W ₁ is measured. As shown in FIG. 28, the threshold value PS is set by adding a prescribed offset value F to this average light quantity E (E + F). As shown in FIG. 27, the regular reflection portion R ₁ and the non-regular reflection portion R ₂ are formed around the welded portion 12, so that the threshold value PS removes the component of the reflected light from the regular reflection portion R _1. It is set to be possible. That is, the offset value F is preset so as to correspond to the amount of light from the specular reflection portion R ₁ . Using this threshold value PS, as shown in FIG. 29, the component of the reflected light from the specular reflection portion R ₁ is removed. That is, the pixels corresponding to the light amount exceeding the threshold value PS are removed from the pixels in the inspection window W ₁ set in the original image. After obtaining the average light intensity G for the remaining pixels, the average light intensity G is compared with a prescribed threshold value QS, and G> QS
If is satisfied, it can be determined that the gloss is insufficient, and thus the welding strength is determined to be poor.

In short, as shown in FIG. 26, after setting the inspection window W ₁ (S1), the inspection window W ₁
The average light amount (average density) E of the above is calculated (S2), and the offset value F is added to the average light amount E to determine the threshold value PS (S3). From the pixels in the inspection windows W ₁ to remove the pixels brighter than the threshold value PS (S4), obtaining an average light intensity (average density) G pixels remaining in the examining window W ₁ (S5). This average light intensity G is defined as a threshold value QS
In comparison, if G> QS is established, it is determined that the gloss is poor (S6).

(Method 3-2) This method is a method of judging that the gloss is a good product if the specular reflection portion R ₁ is present. That is, like the method 3-1, the inspection window W
After setting ₁ , the average light quantity E is obtained, and the offset value F is added to this average light quantity E to set the threshold value PS.
The number H of pixels brighter than the threshold value PS is determined, and when the number H of pixels is larger than a prescribed threshold value RS, that is, when H> RS, the specular reflection portion R ₁ exists in the welded portion 12. Therefore, the gloss is good, and the welding strength is also judged to be a good product. In the case of poor gloss, the relationship shown in FIG. 30 is established, and there are no pixels brighter than the threshold value PS, or even if there are a few pixels, it is possible to easily determine whether the gloss is good or bad.

(Method 3-3) In this method, a search direction is determined using a mask, pixels brighter than the threshold value PS are sequentially tracked, and such pixels are continuously present in the search direction. It is to determine whether or not there is. That is, by introducing the technique of Method 2-3, it is also possible to determine the continuity of pixels satisfying the condition of brightness, and to determine the quality of gloss more accurately as compared with Method 3-2. Is. Method 3-2 for other points
Is the same as.

(Method 3-4) Method 3-1 to Method 3-
In the technique of No. 3, the method of introducing the illumination light to the inspection object 11 is an important element. Therefore, in the present method, the illumination light is switched between the first illumination capable of introducing the specular reflected light and the second illumination capable of introducing the diffuse reflected light to the image input device 1. . By switching between the two illuminations, the quality of the gloss can be determined in consideration of the property of the reflected light, and a more accurate determination can be performed. The other points are the same as those of the methods 3-1 to 3-3.

By the above-mentioned processes, it is possible to judge the presence or absence of the welded portion 12, the appropriateness of the position of the welded portion 12, and the strength of the welded portion 12, and the results of these determinations should be fed back to the welding operation. By this, the quality of welding can be improved. That is,
When a defect is determined in each process, a comprehensive evaluation of the welding state is performed by determining the occurrence rate of each defect for each type. For example, if there are a lot of defects that the welding part 12 does not exist, there is a high possibility that there is an abnormality in the welding device, so a notification is issued to instruct inspection of the welding device. Further, if there are many defects relating to the position of the welded portion 12, it may be instructed to correct the position of the welded portion 12. In this way, by feeding back to the welding conditions and the operation of the welding device using the result of the quality judgment described above,
The yield of welded products can be improved.

[0054]

As described above, according to the present invention, the process of detecting the presence / absence of a welded part, the process of determining whether or not the welded part is located at a required position of the inspection object, and the gloss of the welded part are determined. Since it has a process of judging the welding strength based on, it is possible to make a comprehensive judgment on the welded part,
There is an advantage that an accurate pass / fail judgment can be made regarding welding. As a result, it becomes possible to manage the operation of the welding apparatus and correct the welding position, which has the advantage of improving the quality of the product having the welded portion, that is, improving the yield.

[Brief description of drawings]

FIG. 1 is an explanatory diagram showing a procedure of method 1-1.

FIG. 2 is a diagram showing an example of an inspection object of Method 1-1.

FIG. 3 is a diagram showing a relationship between scan lines and edge flags in method 1-1.

FIG. 4 is a diagram illustrating the concept of a search window in method 1-1.

FIG. 5 is a diagram illustrating a concept of connecting search windows in method 1-1.

FIG. 6 is a block diagram of a processing circuit used in the embodiment.

FIG. 7 is a diagram showing a concept of a local parallel window in the embodiment, (a) is a position in an image, and (b) is a diagram showing a pixel configuration of the local parallel window.

FIG. 8 is a diagram showing a concept of a differential image in the embodiment.

FIG. 9 is an explanatory diagram of a process of obtaining an edge image from a differential image in the embodiment.

FIG. 10 is a diagram showing a relationship among an original image, a differential image, a differential direction image, and an edge image in the example.

FIG. 11 is a diagram illustrating a direction search window used in method 1-2.

FIG. 12 is a diagram showing an example of setting a differential direction value in method 1-2.

FIG. 13 is an explanatory diagram showing a procedure of method 2-1.

FIG. 14 is a diagram showing a concept of Method 2-1.

FIG. 15 is a diagram showing the concept of a criterion in Method 2-1.

FIG. 16 is a diagram showing the concept of a criterion in Method 2-2.

FIG. 17 is a diagram showing a concept of Method 2-3.

FIG. 18 is an explanatory diagram of a connectivity confirmation window used in method 2-3.

FIG. 19 is an explanatory diagram showing a procedure of method 2-4.

FIG. 20 is a diagram showing the concept of Method 2-4.

FIG. 21 is a diagram showing a determination example of method 2-4.

FIG. 22 is an explanatory diagram showing the procedure of method 2-6.

FIG. 23 is a diagram showing the concept of Method 2-6.

FIG. 24 is an explanatory diagram showing the procedure of method 2-7.

FIG. 25 is a diagram showing the concept of method 2-7.

FIG. 26 is an explanatory diagram showing the procedure of method 3-1.

FIG. 27 is a diagram illustrating the principle of method 3-1.

FIG. 28 is a diagram showing a relationship between an average light amount and a threshold value in method 3-1.

FIG. 29 is a diagram showing a result of processing a non-defective product by the method 3-1.

FIG. 30 is a diagram showing a processing result of a defective product by the method 3-2.

[Explanation of symbols]

 1 Image Input Device 2 Analog-Digital Converter 3 Pre-Processing Unit 4 Original Image Memory 5 Differential Image Memory 6 Differential Direction Image Memory 7 Edge Image Memory 8 Arithmetic Processing Unit

Claims (15)

[Claims] 1. A region including a welded portion is imaged by irradiating light on an inspection object having the welded portion so that a difference in illuminance occurs between the welded portion and a peripheral portion of the welded portion. In the weld appearance inspection method, which evaluates the welding state by the appearance of the weld by performing image processing on the image, the process of detecting the presence or absence of a weld and the weld is located at the required position of the inspection object. A welded portion appearance inspection method, comprising: a step of determining whether or not it is, and a step of determining welding strength based on the gloss of a welded portion. 2. A differential image in which the value of each pixel is a differential value corresponding to the rate of change in density with respect to an adjacent pixel, based on an original image which is a grayscale image, and the direction of change in density between adjacent pixels. Create a differential direction image with the differential direction value as the value of each pixel, set an inspection window containing multiple pixels in the image,
The scanning line is moved to scan the pixels in the inspection window to find edge candidate points considered to be the edges of the welded portion based on at least one of the differential value and the differential direction value of each pixel, and then the adjacent edge A candidate point is sequentially obtained based on at least one of a differential value and a differential direction value, and when the number of edge candidate points exceeds a prescribed threshold value, it is determined that a welding site exists in the inspection window. Item 1. A welded part appearance inspection method according to Item 1. 3. A direction search window is further set in the inspection window, pixels in the direction search window are scanned, and a frequency distribution of differential direction values is calculated for pixels whose differential value is greater than or equal to a specified threshold value. 3. The method for inspecting the appearance of a welded part according to claim 2, wherein the scanning line is set along the direction of high frequency and the direction of low frequency is set as the moving direction of the scanning line. 4. A terminal reference line which is an edge of one of the welded members is set, and the total number of pixels between each edge candidate point and the terminal reference line obtained by the method of claim 2 is obtained, and this total number is obtained. The weld appearance inspection method is characterized in that the suitability of the position of the welded portion is determined by comparing with the prescribed threshold value. 5. An area between a welded edge line and a terminal reference line in which a terminal reference line serving as an edge of one welded member is set and edge candidate points obtained by the method of claim 2 are continuous. And the suitability of the position of the welded portion is determined by comparing this area with a prescribed threshold value. 6. A pixel on the edge of one welded member is obtained based on at least one of a differential value and a differential direction value of each pixel by scanning in an inspection window, and a pixel satisfying the same condition is defined. The welded portion appearance inspection method according to claim 4 or 5, wherein when the number of consecutive pixels is equal to or more than one, the consecutive pixels are used as the terminal reference line. 7. A threshold value for density is set based on a tendency of increase or decrease in density of pixels on a scanned line by scanning in a direction substantially orthogonal to an edge of one member welded starting from a prescribed search start point. It is characterized in that a pixel whose density level is inverted with respect to a threshold value on a line set and scanned is used as a terminal point and a straight line passing through the two terminal points obtained by the same procedure is used as a terminal reference line. Item 4. The method for inspecting the appearance of a welded part according to claim 4 or claim 5. 8. The welded portion appearance inspection method according to claim 7, wherein when there is no terminal point on a line scanned from the search start point, a new search is performed by shifting a predetermined pixel in a direction orthogonal to the scanning direction. A method for inspecting the appearance of a weld, which sets a starting point and obtains a terminal point with respect to this search starting point. 9. The total number of pixels between each edge candidate point obtained by the method of claim 2 and a prescribed side of an inspection window set at a fixed position with respect to the inspection object, and the total number is defined. A method for inspecting the appearance of a weld, characterized by determining the suitability of the position of the welded portion by comparing with a threshold value. 10. An area between a weld edge line connecting each edge candidate point obtained by the method of claim 2 and a prescribed side of an inspection window set at a fixed position with respect to an inspection object, A weld appearance inspection method characterized by determining the suitability of the position of a welded portion by comparing this area with a prescribed threshold value. 11. A received light amount corresponding to each pixel is obtained within an inspection window which is set so as to include a welded portion in an original image which is a grayscale image, and a prescribed offset value is added to the average light amount of all pixels within the inspection window. The average light amount is obtained for pixels whose received light amount is smaller than the threshold value, and when the average light amount is larger than a prescribed threshold value, it is determined that the gloss is insufficient and the welding strength is poor.
The weld appearance inspection method described. 12. The amount of received light corresponding to each pixel is determined within an inspection window which is set so as to include a welding portion in an original image which is a grayscale image, and a prescribed offset value is added to the average amount of light of all the pixels within the inspection window. 2. The weld appearance inspection method according to claim 1, wherein when the total number of pixels whose received light amount is larger than the threshold value is larger than the specified threshold value, the weld strength is judged to be good. 13. A mask for defining a search direction is set so as to include a welding part in an original image which is a grayscale image, and pixels having a received light amount equal to or larger than a threshold value are continuously present in a specified number or more in the search direction. The weld appearance inspection method according to claim 1, wherein the weld strength is determined to be good. 14. A light source for which specular reflection light is imaged and a light source for which diffuse reflection light is imaged are turned on at different timings, and the quality of the gloss of the welded part is determined based on the light from each light source. Claim 11 thru | or Claim 13 characterized by the above-mentioned.
The visual inspection method for welded parts according to any one of 1. 15. The operation of the welding apparatus is corrected by using at least one of a determination result of presence / absence of a welding site, a determination result of suitability of a position of the welding site, and a determination result of welding strength based on gloss of the welding site. The welding portion appearance inspection method according to claim 1, wherein the welding position is corrected.