JP2003311464A - Instrument and method for inspecting welded state in annular welded part - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain an instrument and a method for inspecting welded state in an annular welded part with which the inspection of a portion welded as the annular shape on a metal plate, etc., can quickly be performed over the whole zone of the welded part containing the welded state regardless of the surface state of the metal plate. <P>SOLUTION: This inspecting instrument is provided with a lighting means 2 disposed so as to apply the light in the vertical direction to the metal surface 11 containing the annular shape welded part on the material 1 to be inspected, an image pickup means 3 for receiving the reflected light reflected with the metal surface 11 applied from the lighting means 2 and a welded-state judging means 9 for judging normal or defective welded state from image data of the annular welded part picked up with an image pickup means 3. This inspecting method is performed with which the picked up image data 31 are emphasized and welded positional coordinates 35 of the whole circle of the annular welded part is detected from the emphasized image data 32, and a welding reference circle 36 is calculated from the detected welded positional coordinates 35 and the normal or defective welded state, is judged by comparing the detected welded positional coordinates 35 and the calculated welding reference circle 36. <P>COPYRIGHT: (C)2004,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention automatically checks the welding condition of metal parts having ring-shaped welded portions such as oil dashpots used as timed elements in electromagnetic trip devices for circuit breakers. The present invention relates to a welding state inspection device and inspection method for a ring-shaped welded portion.

[0002]

2. Description of the Related Art As a conventional inspection device for a welded part of a metal part, a welded part of a part to be inspected made of a cylindrical metal part is irradiated with light obliquely from a direction along a cutting direction. , The reflected light is imaged by a television camera or the like arranged above the welded portion, and the density (shadeness) average of the position on the welding line in the imaged two-dimensional image and the partial area at the position other than the welding line is obtained, A method of determining an unwelded portion based on the magnitude of the difference is proposed. In addition, in Japanese Unexamined Patent Publication No. 8-267269, in order to determine an unwelded portion, in addition to the direction along the cutting direction, a two-dimensional image is obtained by irradiating light from a direction perpendicular to the cutting direction and obliquely. There is disclosed a method of inspecting a welded portion by using images also without being influenced by the surface condition (presence or absence of burning) of a metal component which is an inspection target component.

[0003]

Since the conventional weld inspection device is constructed as described above, since the weld inspection is performed locally, it is necessary to inspect all parts of the weld. It requires a lot of inspection time, and further, in the conventional inspection, there is a problem that only the presence / absence of the welding is inspected, and the inspection of the welding state such as the dispersion of welding cannot be performed.

The present invention has been made in order to solve the above problems, and the inspection of a portion welded in a ring shape to a metal plate or the like is not affected by the surface state of the metal plate, and the welding state is not affected. An object of the present invention is to provide a welding state inspection device and an inspection method for a ring-shaped welded portion that can be rapidly performed over the entire welded portion.

[0005]

A welding state inspection device for a ring-shaped welded portion according to the present invention is a lighting means arranged so as to irradiate light perpendicularly to a ring-shaped welded portion forming surface of an object to be inspected. And the ring-shaped welded portion of the object to be inspected, the ring-shaped welded portion of the object to be inspected is received, and the reflected light emitted from the illumination means and reflected by the surface of the ring-shaped welded portion is received. An image pickup means for picking up an image of the entire area at the same time, and a welding state judging means for judging whether the welding state is good or bad from the image data of the ring-shaped welded portion picked up by the image pickup means are provided.

Further, the welding state determining means is composed of an electronic computer, and an image enhancing means for performing an image enhancing process on the image data of the ring-shaped welded portion imaged by the image capturing means, and an image enhancing means. Welding position detection means for detecting a predetermined number of welding positions on the entire circumference of the ring-shaped welded portion from the image data emphasized by, and from the welding position of the ring-shaped welded portion detected by the welding position detection means to the ring-shaped welded portion Of the welding state of the ring-shaped welded portion by comparing the welding reference circle calculating means for calculating the welding reference circle with the welding position detected by the welding position detecting means and the welding reference circle calculated by the welding reference circle calculating means. It is provided with a welding determination means for determining good / bad.

The image enhancing means is for subjecting the image data of the ring-shaped welded portion to shrinkage filter processing. Further, the welding position detecting means, in the image data obtained by imaging the ring-shaped welded portion, has density values on a straight line radially extending from the center of the ring-shaped welded portion, where the density values are from bright (white) to dark (black) and dark (black). ) To bright (white), two edge points are obtained, and the middle point of the two edge points is detected as the welding position.

Further, the welding reference circle calculating means calculates L based on the welding position of the ring-shaped welded portion detected by the welding position detecting means.
The welding standard circle is calculated using a MedS (Least Median of Squares) robust circle regression algorithm. Further, the welding determination means determines the difference between the radius of the welding reference circle obtained by the welding reference circle calculation means and the distance from the center of the welding reference circle to each welding position of the ring-shaped welded portion along the entire circumference of the ring-shaped welded portion. And the standard deviation value is compared with a preset threshold value to determine whether the welding state of the ring-shaped weld is good or bad. In addition, the welding determination means obtains the distance from the center of the welding reference circle obtained by the welding reference circle calculation means to each welding position of the ring-shaped welded portion, and the difference in the distance between the adjacent welding positions is ring-shaped welded. The goodness / badness of the welding state of the ring-shaped welded portion is determined by calculating the standard deviation value over the entire circumference of the welded portion and comparing the standard deviation value with a preset threshold value.

Further, according to the method for inspecting the welded state of the ring-shaped welded portion according to the present invention, the reflected light that is irradiated perpendicularly to the ring-shaped welded portion forming surface of the object to be inspected and reflected by the ring-shaped welded portion forming surface Of the ring-shaped welded portion imaged by the image pickup means at the same time by receiving an image of the entire area of the ring-shaped welded portion of the inspection object by receiving light by the image pickup means arranged to face the ring-shaped welded portion forming surface. It includes a welding state determination step of determining whether the welding state is good or bad from the image data.

Further, in the welding state determining step, an image enhancing step of performing an image enhancing process on the image data of the ring-shaped welded portion captured by the image capturing means, and an image enhanced by the image enhancing step. Welding position detection process that detects a predetermined number of welding positions around the entire circumference of the ring-shaped welded part from the data, and the welding reference circle of the ring-shaped welded part is calculated from the welding position of the ring-shaped welded part detected in the welding position detection process The welding reference circle calculation process to be performed and the welding position detected in the welding position detection process and the welding reference circle calculated in the welding reference circle calculation process are compared to determine whether the welding state of the ring-shaped weld is good or bad. And a welding determination step.

In the image enhancement step, shrinkage filter processing is performed on the image data of the ring-shaped welded portion. In addition, in the welding position detection step, in the image data of the ring-shaped welded portion, the density values are light (white) to dark (black) and dark (black) on a straight line radially extending from the center of the ring-shaped welded portion. ) To bright (white), two edge points are obtained, and the middle point of the two edge points is detected as the welding position. Also, in the welding position detection process, the density value changes from bright (white) to dark (black) and from dark (black) to bright (white) on a straight line extending radially from the center of the ring-shaped weld. When the individual edge points are not obtained, it is determined that the welded portion on the straight line is unwelded.

Further, in the welding reference circle calculating step, L based on the welding position of the ring-shaped welded portion detected in the welding position detecting step.
The welding standard circle is calculated using a MedS (Least Median of Squares) robust circle regression algorithm. In addition, in the welding determination step, the difference between the radius of the welding reference circle obtained in the welding reference circle calculation step and the distance from the center of the welding reference circle to each welding position of the ring-shaped weld is determined by measuring the entire circumference of the ring-shaped weld. Is calculated, and the standard deviation value is compared with a preset threshold value to determine whether the welding state in the ring-shaped welded portion is good or bad. In the welding determination step, the distance from the center of the welding reference circle obtained in the welding reference circle calculation step to each welding position of the ring-shaped weld is calculated, and the difference in the distance between adjacent welding positions is ring-shaped welded. It is calculated for the entire circumference of the part and the standard deviation value is compared with a preset threshold value to judge the good / bad of the welding state in the ring-shaped welded part.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiment 1. Hereinafter, a welding state inspection device and inspection method for a ring-shaped welded portion, which is an embodiment of the present invention, will be described with reference to the drawings. 1 is a block diagram showing a configuration of a welding state inspection device for a ring-shaped welded portion according to a first embodiment of the present invention, and FIG. 2 shows a flow of operations of the welding state inspection device for a ring-shaped welded portion according to the first embodiment. It is a flowchart.

In the figure, 1 is an object to be inspected which is made of a metal part and has a ring-shaped welded portion, and 2 is perpendicular to a ring-shaped welded portion forming surface (hereinafter abbreviated as ring-shaped welded surface) of the object 1 to be inspected. Illuminating means arranged to irradiate light on the
ED lighting or the like is used. 3 is arranged to face the ring-shaped welded surface of the inspection object 1, receives the illumination light (reflected light) reflected from the surface of the inspection object 1 including the welded portion, and An image pickup device for generating image data, for example, a CCD (electrically coupled device; Char) is used as an image pickup element.
For example, a television camera by Ge Coupled Device) is used. Reference numeral 4 denotes a welding state determination means which is an electronic computer which receives the grayscale image data of the welded portion from the image pickup means 3 and determines the welding state of the ring-shaped welded portion, and includes an image data storage means 5, an image enhancement means 6, A welding position detecting means 7, a welding reference circle calculating means 8 and a welding determining means 9 are included.
The illumination unit 2 includes a control unit (not shown) that controls the illumination intensity, and the illumination unit 2 and the image pickup unit 3 are provided.
Is provided with position adjusting means (not shown) for adjusting the respective distances of the inspected object 1, the illuminating means 2, and the imaging means 3.

The image data storage means 5 is composed of a storage medium such as a hard disk device of an electronic computer or a CD-ROM or a DVD which constitutes the welding state inspection means 4, and stores the image data input from the image pickup means 3. .
The image emphasizing unit 6 applies contraction filter processing as an emphasizing process to the image data generated by the imaging unit 3.
The welding position detecting means 7 obtains the position coordinates of the ring-shaped welded portion from the image data subjected to the contraction filter processing by the image enhancing means 6. The welding reference circle calculation means 8 obtains a welding reference circle from all position coordinates of the welded portion detected by the welding position detection means 7. The welding determination means 9 compares the welding reference circle obtained by the welding reference circle calculation means 8 with the designed ring-shaped welded portion to determine whether the welding state is good or bad.

Next, a welding state inspection method using the welding state inspection apparatus according to this embodiment will be described with reference to the flow chart shown in FIG. First, as shown in FIG. 1, light can be emitted perpendicularly to the ring-shaped welded surface of the inspection object 1, and the entire area of the ring-shaped welded portion should be within the field of view of the CCD camera which is the imaging means 3. The imaging means 3, the illuminating means 2, and the inspection object 1 are arranged in the same, the inspection object 1 is irradiated with light by the illumination means 2, and the reflected light is received by the imaging means 3 to generate grayscale image data. (ST1).

At this time, the light radiated perpendicularly to the object 1 to be inspected by the illuminating means 2 is, as shown in FIG. 3 (a), on the metal surface 11 of the metal part constituting the object 1 to be inspected. Although the light is specularly reflected (reflected light 21), in the welded portion 12, the reflectance is lower than that of the metal surface 11 and the reflected light 21 is less than that of the metal surface 11 due to the burning that occurs during welding. Therefore, as shown in FIG. 3B, in the image data 31 generated by the image pickup means 3 by receiving the reflected light 21, the welded portion 12 is imaged in black. The reflected light 21 received by the imaging means 3
Position adjusting means (not shown) so that the strength of the
May be used to adjust the respective distances of the inspection object 1, the illumination means 2, and the imaging means 3.

Next, the image data 31 generated by the image pickup means 3 is immediately input to the welding state determination means 4 and stored in the image data storage means 5 in the welding state determination means 4 (ST2).

Next, the image data 31 stored in the image data storage means 5 is read out, and in the image enhancing means 6,
After the contraction filter process, which is the emphasis process for the image data 31, is performed several times, the image data 32 subjected to the contraction filter process is stored again in the image data storage means 5 (ST3). FIG. 4 is a diagram for explaining the contraction filter process for the image data in the image emphasizing means.
(A) shows the change of the image data by the contraction filter process, and (b) shows the change at the pixel level by the contraction filter process. In FIG. 4, 31a is a local image showing the pixel level state of the image data 31 before contraction filter processing, and 32a is a local image showing the pixel level state of the image data 32 after contraction filter processing. The contraction filter process is a process in which a minimum pixel value is taken in a rectangular area (for example, 3 vertical pixels × 3 horizontal pixels) including a point of interest in the image data, and this is set as the pixel value of the point of interest.

Specifically, in the local image 31a before the contraction filter processing, the pixel value of the black-painted portion is the lowest,
It is shown that the pixel value of the white-painted part is the highest, and the shaded part is higher than the pixel value of the black-painted part and lower than the pixel value of the white-painted part. When contraction filter processing is applied to the local image 31a, the local image 32 in which the pixels adjacent to the black-painted portion are replaced with the black-painted portion having the lowest pixel value.
Change to a. By performing the contraction filtering process on the entire image, as shown in FIG. 4A, the image data 32 in which the welded portion is emphasized is obtained. In the present embodiment, an example in which contraction filter processing is performed as the enhancement processing for the image data 31 has been shown, but other image processing methods that can enhance the image data including the captured welded portion may be used. It is not limited to the contraction filtering method.

Next, the enhanced image data 32 stored in the image data storage means 5 is read out, and the welding position detecting means 7 detects the welding position coordinates from the image data 32 (ST4). FIG. 5 is a diagram for explaining the process of detecting the welding position coordinates from the image data in the welding position detecting means, FIG. 5A is the image data after the emphasizing process read from the image data storage means, and FIG. [Fig. 3] is a partially enlarged view of (a). In FIG. 5, reference numeral 33 is an edge detection straight line for finding the edge points of both ends of the welded portion from the image data 32, and a total of 360 radial lines are provided at 1 ° intervals from the center position of the ring-shaped welded portion. Welding position coordinates are edge detection line 3
3 using a one-dimensional differential filter, an edge point 34a where the density value changes from light (white) to dark (black) and an edge point 34b where the density value changes from dark (black) to light (white) are obtained. It is detected by setting the midpoint coordinates of the point 34a and the edge point 34b as the welding position coordinates. The same processing is applied to 360 edge detection straight lines 33 to obtain welding position coordinates of 360 points in total.

Next, the welding reference circle calculating means 8 obtains a welding reference circle from the welding position coordinates of 360 points detected by the welding position detecting means 7 (ST5). FIG. 6 is a diagram for explaining the process of calculating the welding reference circle from the welding position coordinates in the welding reference circle calculation means 8, and FIG. 6A is a diagram showing a group of welding position coordinates detected by the welding position detection means, FIG. 6B is a diagram showing a welding reference circle obtained from the welding position coordinate group. In the figure, 35 is a welding position coordinate (total 360 points) detected by the welding position detecting means 7, and 36 is a welding reference circle calculated from the total 360 welding position coordinates 35.

The center coordinates (a, b) of the welding reference circle 36 and the radius r are based on the welding position coordinates 35 of 360 points in total.
It is calculated from the equation (1) indicating the welding reference circle 36 using the LMedS robust circle regression algorithm described later. (X−a) ² + (y−b) ² −r ² = 0 ... (1)

In the LMedS robust circular regression algorithm, the LMedS (Least Median of Squares) criterion is one of the robust fitting criteria and was proposed by Rousseeuw (reference: Rousseeuw RJ
, Leroy AM "Robust Regression and Outlier Detec
tion "in John Wiley & Sons 1986 etc.). LMed
The S criterion is represented by the following equation (2) and is a criterion for minimizing the median of the square of the deviation ri. LMedS = min median (ri) ² ... (2)

On the other hand, the least-squares method (LMS = Least Minimum of Squares) which is the most general fitting criterion is expressed by the following equation (3). LMS = min Σ (ri) ² (3)

The major feature of the LMedS standard is the outlier.
(Outlier), which is an exceptional value that jumps out of the measured values, is automatically excluded from the calculation, so that a stable result is always obtained. Theoretically,
If the outlier is less than 50% of the measured value, the correct value can be estimated even if the outlier is included. on the other hand,
According to the fitting criterion using the least squares method, if there is at least one outlier in the measured values, the regression parameters (center coordinates (a, b), radius r) are calculated including this value (outlier). Therefore, it may not be possible to estimate the correct value.

Next, from the welding position coordinates 35 detected by the welding position detecting means 7 by the welding determining means 9 and the welding reference circle 36 calculated by the welding reference circle calculating means 8, the welding reference circle 36 is obtained.
The deviation (error) of each welding position coordinate 35 with respect to is obtained to determine whether the welding state is good or bad (ST6). FIG. 7 is a diagram for explaining the welding determination means. In the figure, 35 is the welding position coordinate detected by the welding position detecting means 7 (total 3
60 points), 36 are welding reference circles calculated by the welding reference circle calculating means 8 based on the welding position coordinates 35 of 360 points in total, r is the radius of the welding reference circle 36, and (a, b) are welding reference circles 36. The center coordinate of the, dis is the welding reference circle 3 of each welding position coordinate 35
The distance from the center coordinates (a, b) of 6 is shown.

Next, a method of judging the welding state by the welding judging means 9 will be described. First, a total of 360 welding position coordinates 35 (coordinate position: x detected by the welding position detecting means 7).
For each of (i) and y (i), the following equation (4)
Is used to determine the distance dis (i) from the center coordinates (a, b) of the welding reference circle 36. However, i = 0, 1, ...
-359. dis (i) = {(x (i) -a) ² + (y (i) -b) ² } ^1/2 ... (4)

Next, the radius of the radius r of the welding reference circle 36 and the distance dis (i) from the center coordinates (a, b) of the welding reference circle 36 obtained for the welding position coordinates 35 of 360 points in total. The error rdiff (i) is calculated using the following equation (5). rdiff (i) = dis (i) −r ... (5)

Next, the standard deviation value std from the radius error rdiff (i) at the welding position coordinates 35 of 360 points in total.
ev is obtained, and the ratio (%) of the obtained standard deviation value stdev to the radius r of the welding reference circle 36 is used as a judgment value JudgeVal of good / bad welding state, and the following equation (6) is obtained.
Calculate using. JudgeVal = (stdev / r) × 100 ... (6)

Next, the judgment value Ju obtained by using the equation (6)
dgeVal is a preset threshold ThJudge
It is checked whether or not the following conditional expression (7) is satisfied with respect to Val. JudgeVal ≦ ThJudgeVal ··· (7)

Judgment value Judge obtained using equation (6)
If Val satisfies the conditional expression (7), it is determined that the welding state is good, and if not, it indicates that the welding position coordinate 35 is largely deviated from the welding reference circle 36. As shown in FIG. 8, there is a possibility that welding dust 37 or the like has occurred, and the welding state is determined to be poor.

As described above, according to the present embodiment, the illuminating means 2 is arranged so that the metal surface 11 including the ring-shaped welded portion of the inspection object 1 can be vertically irradiated with light, and the illuminating means 2 is arranged. The reflected light reflected from the metal surface 11 including the welded portion from 2 is received by the imaging means 3 to image the entire area of the ring-shaped welded portion, and the captured image of the welded portion is emphasized, and then the welding position coordinate 35. Is determined to determine the welding reference circle 36, and whether the welding state is good or bad is determined from the variation of the welding position coordinates 35 with respect to the welding reference circle 36. Therefore, the welding state is not affected by the surface state of the metal plate and the ring-shaped welded portion is obtained. It is possible to quickly inspect the welding state for the entire area. Further, the welding state inspection device for the ring-shaped welded portion according to the present embodiment can be realized with a relatively simple device configuration.

Embodiment 2. In the first embodiment, the radius error rdiff between the welding reference circle 36 and the welding position coordinate 35.
The standard deviation value stdev is calculated from (i), and the standard deviation value s
Ratio (%) of tdev to the radius r of the welding reference circle 36
Was used as the judgment value of good / bad of the welding state, but the same effect as that of the first embodiment can be obtained by judging the good / bad of the welding state using the judgment value obtained by the method described below. .

For example, as described in the section of the welding determination means 9 of the first embodiment, the welding position coordinate 35 of 360 points in total is used.
For each of (coordinate position: x (i), y (i)), the distance dis (i) from the center coordinates (a, b) of the welding reference circle 36 is obtained using the equation (4), The distance difference ddiff (i) between adjacent welding positions is calculated using the following equation (8). However, i = 0, 1, ..., 358. ddiff (i) = dis (i) -dis (i + 1) ... (8)

Next, the distance difference dd between the adjacent welding positions.
If the standard deviation value stdev2 is obtained from the if (i), the ratio (%) of the obtained standard deviation value stdev2 to the radius r of the welding reference circle is used as the judgment value Judge of the welding condition.
It is calculated as Val2 using the following equation (9). JudgeVal2 = (stdev2 / r) × 100 ... (9)

Next, the judgment value Ju obtained by using the equation (9)
dgeVal2 is a preset threshold ThJudg
It is checked whether eVal2 satisfies the following conditional expression (10). JudgeVal2 ≤ ThJudgeVal2 (10)

Judgment value Judge obtained using equation (9)
If Val2 satisfies the conditional expression (10), the welding state is determined to be good, and if not, it indicates that the welded portion is rough in a local range, and the welding state is poor. Is determined.

Embodiment 3. In the first and second embodiments, the error between the welding reference circle 36 and the welding position coordinate 35 is checked to detect a defective welding state such as welding scatter. However, in the welding position detecting means 7 shown in the first embodiment, When the welding position is obtained by detecting the edge point where the density value changes from bright (white) to dark (black) and the edge point where the density value changes from dark (black) to bright (white) from the image data 32, FIG. As shown in FIG. 5, since the edge point (welding position) cannot be detected in the unwelded portion 38, the presence / absence of welding can be inspected at the same time. At this time, the position of the unwelded portion can be specified together with the presence or absence of welding.

[0040]

As described above, according to the present invention, the welding state inspection device for the ring-shaped welded portion is arranged so as to irradiate light perpendicularly to the ring-shaped welded surface of the inspection object. And the entire area of the ring-shaped welded part of the object to be inspected at the same time by receiving the light reflected from the ring-shaped welded surface that is radiated from the illumination means Image forming means and a welding state determining means for determining whether the welding state is good or bad based on the image information of the ring-shaped welded portion picked up by the image pickup means. The inspection can be performed with a relatively simple device configuration.

Further, after enhancing the image of the picked up welded portion, a predetermined number of welding positions are detected on the entire circumference of the ring-shaped welded portion to obtain a welding reference circle, and the detected welding position and the welding reference circle are compared. In order to determine whether the welding state is good or bad, it is possible to reduce the influence of the surface state of the metal surface including the ring-shaped welded portion of the inspection object and perform an accurate determination,
Since the processing is performed in the electronic computer, the welding state can be quickly inspected over the entire area of the ring-shaped welded portion.

Further, by applying contraction filter processing as emphasis processing to the image information obtained by imaging the ring-shaped welded portion,
It is possible to absorb the influence on the reflected light due to a slight crack or stain on the welded portion, and it is possible to reduce erroneous determination.

With respect to the image information subjected to the image enhancement processing, the density values are bright (white) to dark (black) and dark (black) on a straight line radially extending from the center of the ring-shaped weld. The two edge points that change from light to bright (white) are detected and the midpoint is detected as the welding position, so the influence of the material of the metal that constitutes the object to be inspected (ring-shaped weld) Without it, the welding position can be specified accurately.

Further, the reference circle of the ring-shaped welded portion is detected from the welding positions on the entire circumference of the ring-shaped welded portion detected by the welding position detecting means by using the LMedS (Least Median of Squares) robust circle regression algorithm. By calculating, the exceptional values that are far out of the range of the detected values of the welding position are automatically excluded from the calculation, so that an accurate welding reference circle can be calculated.

Further, the difference between the radius of the welding reference circle and the distance from the center of the welding reference circle to each welding position of the ring-shaped weld is calculated for the entire circumference of the ring-shaped weld, and the standard deviation value is used to perform welding. In order to determine whether the state is good or bad, it is possible to quantitatively evaluate the circularity of the ring-shaped welded portion and determine whether the state of welding is good or bad.

Further, the distance from the center of the welding reference circle to each welding position of the ring-shaped welded portion is obtained, and the difference in distance between the adjacent welding positions is calculated with respect to the entire circumference of the ring-shaped welded portion, and the standard is calculated. Since the goodness / badness of the welding state is determined from the deviation value, the local welding roughness of the ring-shaped welded portion can be quantitatively evaluated to determine the goodness / badness of the welding state.

Further, in the welding state inspection of the ring-shaped welded portion, light is radiated perpendicularly to the ring-shaped welded surface of the inspection object, and the regular reflection light is received to detect the entire area of the ring-shaped welded portion of the inspection object. Are simultaneously imaged, and the goodness / defectiveness of the welding state is determined from the imaged image information. Therefore, the welding state in the entire area of the ring-shaped welded portion can be inspected by one imaging operation.

Further, after enhancing the image of the captured welded portion, a predetermined number of welding positions are detected on the entire circumference of the ring-shaped welded portion to obtain a welding reference circle, and the detected welding position and the welding reference circle are compared. Since the goodness / badness of the welding state is determined, the influence of the surface state of the metal surface including the ring-shaped welded portion of the inspection object can be reduced and an accurate determination can be performed.

Further, by emphasizing the image information of the ring-shaped welded portion by shrinkage filter processing, it is possible to absorb the influence on the reflected light due to the minute crossing or dirt of the welded portion, and the erroneous determination is made. Can be reduced.

With respect to the image information subjected to the image enhancement processing, the density values are bright (white) to dark (black) and dark (black) on a straight line extending radially from the center of the ring-shaped weld. The two edge points that change from light to bright (white) are detected and the middle point is detected as the welding position, so the influence of the material of the metal that constitutes the object to be inspected (ring-shaped weld) Without it, the welding position can be specified accurately.

With respect to the image information subjected to the image enhancement processing, the density values are bright (white) to dark (black) and dark (black) on a straight line radially extending from the center of the ring-shaped weld. If two edge points that change from light to bright (white) are not obtained, it is possible to determine that the welded portion on the straight line is an unwelded portion,
The position of the unwelded portion can be specified together with the presence or absence of welding.

Further, the reference circle of the ring-shaped welded portion is detected from the welding positions on the entire circumference of the ring-shaped welded portion detected by the welding position detecting means by using the LMedS (Least Median of Squares) robust circle regression algorithm. By calculating, the exceptional values that are far out of the range of the detected values of the welding position are automatically excluded from the calculation, so that an accurate welding reference circle can be calculated.

Further, the difference between the radius of the welding reference circle and the distance from the center of the welding reference circle to each welding position of the ring-shaped weld is calculated for the entire circumference of the ring-shaped weld, and the standard deviation value is used to perform welding. In order to determine whether the state is good or bad, it is possible to quantitatively evaluate the circularity of the ring-shaped welded portion and determine whether the state of welding is good or bad.

Further, the distance from the center of the welding reference circle to each welding position of the ring-shaped welded portion is obtained, and the difference in the distance between the adjacent welding positions is calculated with respect to the entire circumference of the ring-shaped welded portion, and the standard is calculated. Since the goodness / badness of the welding state is determined from the deviation value, the local welding roughness of the ring-shaped welded portion can be quantitatively evaluated to determine the goodness / badness of the welding state.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a welding state inspection device for a ring-shaped welded portion according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing the flow of the operation of the welding state inspection device for ring-shaped welded parts according to the first embodiment of the present invention.

FIG. 3 is a diagram for explaining an image pickup means in the welding state inspection device for the ring-shaped welded portion according to the first embodiment of the present invention.

FIG. 4 is a diagram for explaining image enhancing means in the welding state inspection device for ring-shaped welded portions according to the first embodiment of the present invention.

FIG. 5 is a diagram for explaining welding position detecting means in the welding state inspection device for the ring-shaped welded portion according to the first embodiment of the present invention.

FIG. 6 is a diagram for explaining a welding reference circle calculation means in the welding state inspection device for the ring-shaped welded portion according to the first embodiment of the present invention.

FIG. 7 is a diagram for explaining welding determination means in the welding state inspection device for the ring-shaped welded portion according to the first embodiment of the present invention.

FIG. 8 is a diagram showing an example of defective welding of a ring-shaped welded portion.

FIG. 9 is a diagram showing another example of defective welding of a ring-shaped welded portion.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 inspection object, 2 illumination means, 3 imaging means, 4 welding state determination means, 5 image data storage means, 6 image enhancement means, 7 welding position detection means, 8 welding reference circle calculation means, 9 welding determination means, 11 Metal surface, 12 welds, 31 image data, 32 image data (after image enhancement processing), 33 edge detection curve local image, 34a, 3
4b Edge point, 35 Welding position coordinate, 36 Welding reference circle, 37 Welding spot, 38 Unwelded part.

─────────────────────────────────────────────────── ─── Continuation of front page (51) Int.Cl. ⁷ Identification code FI theme code (reference) G06T 7/60 200 G06T 7/60 200C (72) Inventor Koichi Sugimoto 2-3-2 Marunouchi, Chiyoda-ku, Tokyo No. Sanryo Electric Co., Ltd. (72) Inventor Naomichi Yamada 2-3-3 Marunouchi, Chiyoda-ku, Tokyo F-Term within Sanryo Electric Co., Ltd. (reference) 2F065 AA06 AA12 AA15 AA17 CC15 DD04 DD06 FF04 GG07 HH13 JJ26 QQ08 QQ18 QQ24 QQ25 QQ32 QQ33 RR00 TT03 2G051 AA90 AB13 AB20 CA04 CB01 EA08 EA12 EC02 ED23 5B057 AA17 BA02 CE03 CE06 DA03 DA07 DB02 DB05 DB09 DC03 DC05 DC09 DC16 5L096 AA03 AA06 BA03 CA02 EA02 EA27 FA04 FA06 FA33 FA62 FA64 FA66 FA69 GA02 GA12 GA51 GA55

Claims (15)

[Claims] 1. An illumination means arranged to irradiate light perpendicularly to a ring-shaped welded portion forming surface of the inspection object, and an illuminating means arranged so as to face the ring-shaped welding portion forming surface of the inspection object. An image pickup means for simultaneously picking up an image of the entire ring-shaped welded portion of the object to be inspected by receiving the reflected light emitted from the illumination means and reflected by the ring-shaped welded portion forming surface; A welding state inspection device for a ring-shaped welded portion, further comprising: welding state determination means for determining whether the welding state is good or bad from the image data of the ring-shaped welded portion. 2. The welding state determination means comprises an electronic computer, and image enhancement means for performing image enhancement processing on image data of the entire ring-shaped welded portion imaged by the imaging means, Welding position detecting means for detecting a predetermined number of welding positions with respect to the entire circumference of the ring-shaped welded portion from the image data emphasized by the image enhancing means, and the ring-shaped welding detected by the welding position detecting means. Welding reference circle calculating means for calculating the welding reference circle of the ring-shaped welded portion from the welding position of the portion, the welding position detected by the welding position detecting means, and the welding reference circle calculated by the welding reference circle calculating means The welding condition inspection device for the ring-shaped welded part according to claim 1, further comprising: a welding determination means for determining whether the welded condition of the ring-shaped welded part is good or bad. 3. The welded state inspection device for a ring-shaped welded portion according to claim 2, wherein the image enhancing means performs shrinkage filter processing on the image data of the ring-shaped welded portion. 4. The welding position detecting means, in the image data obtained by picking up the ring-shaped welded portion, has a density value from bright (white) to dark (black) on a straight line radially extending from the center of the ring-shaped welded portion. , And two edge points that change from dark (black) to bright (white), and the midpoint of the two edge points is detected as the welding position. Welding condition inspection device for welded parts. 5. The welding reference circle calculating means, based on the welding position of the ring-shaped welded portion detected by the welding position detecting means, uses a LMedS (Least Median of Squares) robust circle regression algorithm. The welding condition inspection device for a ring-shaped welded portion according to claim 2, wherein the welding reference circle is calculated using the welding reference circle. 6. The welding determination means determines the difference between the radius of the welding reference circle obtained by the welding reference circle calculation means and the distance from the center of the welding reference circle to each welding position of the ring-shaped welded portion. Is determined for the entire circumference of the ring-shaped welded portion, and the standard deviation value is compared with a preset threshold value to determine whether the welded state in the ring-shaped welded portion is good or bad. 2. The welding state inspection device for ring-shaped welded parts according to 2. 7. The welding determination means obtains the distance from the center of the welding reference circle obtained by the welding reference circle calculation means to each welding position of the ring-shaped welded portion, and the welding position between adjacent welding positions is determined. A feature is that the difference in distance is calculated for the entire circumference of the ring-shaped welded portion, and the standard deviation value is compared with a preset threshold value to judge whether the welded state in the ring-shaped welded portion is good or bad. The welding state inspection device for the ring-shaped welded portion according to claim 2. 8. The reflected light, which is radiated perpendicularly to the ring-shaped welded portion forming surface of the inspection object and is reflected by the ring-shaped welded portion forming surface, is arranged so as to face the ring-shaped welded portion forming surface. The step of simultaneously imaging the entire area of the ring-shaped welded portion of the inspection object by receiving light by the image pickup means, and the good / defective welding state from the image data of the ring-shaped welded portion picked up by the image pickup means. A method for inspecting a welded state of a ring-shaped welded portion, which comprises a step of determining a welded state. 9. The welding state determining step includes an image enhancing step of performing an image enhancing process on the image data of the ring-shaped welded portion captured by the image capturing means, and the image enhancing step. A welding position detection step of detecting a predetermined number of welding positions on the entire circumference of the ring-shaped welded portion from the imaged image data; and a ring shape from the welding position of the ring-shaped welded portion detected in the welding position detection step. The welding reference circle calculation step of calculating the welding reference circle of the welded portion, the welding position detected in the welding position detection step and the welding reference circle calculated in the welding reference circle calculation step are compared, and the ring-shaped welding is performed. The welding state inspection method for a ring-shaped welded portion according to claim 8, further comprising a welding determination step of determining whether the welded state of the portion is good or bad. 10. The method for inspecting a welded state of a ring-shaped welded portion according to claim 9, wherein in the image enhancement step, shrinkage filter processing is applied to image data of the ring-shaped welded portion. 11. In the welding position detecting step, in the image data obtained by imaging the ring-shaped welded portion, the density value is light (white) to dark (black) on a straight line radially extending from the center of the ring-shaped welded portion. , And the two edge points that change from dark (black) to bright (white), and the midpoint of the two edge points is detected as the welding position. Weld condition inspection method. 12. In the welding position detecting step, the density values are bright (white) to dark (black) and dark (black) to bright (white) on a straight line radially extending from the center of the ring-shaped weld. 12. The method for inspecting a welded state of a ring-shaped welded portion according to claim 11, wherein the welded portion on the straight line is determined to be unwelded when two edge points that change to are not obtained. 13. The welding reference circle calculation step uses a LMedS (Least Median of Squares) robust circle regression algorithm based on the welding position of the ring-shaped weld detected in the welding position detection step. The welding state inspection method for a ring-shaped welded portion according to claim 9, wherein the welding reference circle is calculated using the welding reference circle. 14. The welding determination step includes a difference between a radius of the welding reference circle obtained in the welding reference circle calculation step and a distance from a center of the welding reference circle to each welding position of the ring-shaped weld portion. Is determined for the entire circumference of the ring-shaped welded portion, and the standard deviation value is compared with a preset threshold value to determine whether the welded state in the ring-shaped welded portion is good or bad. 9. The method for inspecting the welded state of the ring-shaped welded part according to 9. 15. The welding determination step obtains the distance from the center of the welding reference circle obtained in the welding reference circle calculation step to each welding position of the ring-shaped welded portion, and determines the distance between the adjacent welding positions. A feature is that the difference in distance is calculated for the entire circumference of the ring-shaped welded portion, and the standard deviation value is compared with a preset threshold value to judge whether the welded state in the ring-shaped welded portion is good or bad. The method for inspecting a welded state of a ring-shaped welded portion according to claim 9.