JPH06328251A - Joint detecting method - Google Patents

Abstract

PURPOSE:To provide a joint detecting method capable of detecting the position of joint line with identifying for a joint line or tack welded part when tack welded part exists in the course of welding joint. CONSTITUTION:When a tack welded part 23 exists in the course of a joint line 22 to be welded, the detected point 31 detected at a tack welded part 23 by the surface ruggedness of a tack welded part 23 is found in scattered dot state, when a linear regression remainder value is larger than a decision threshold value, the detecting point is recognized as a tack welded part 23, when a linear regression remainder value is smaller than the decision threshold value, the detected point is recognized as a joint line 22. As a result, by identifying for a joint line 22 or tack welded part 23, the joint line 22 is detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting a seam of a welded joint in, for example, a welding robot or an automatic welding machine, and more particularly to a seam detection method capable of discriminating a seam from a tack portion.

[0002]

2. Description of the Related Art For example, Japanese Patent Publication No. 4-5924 discloses a technique for detecting a joint position of a welded joint using a visual sensor in a welding robot or an automatic welding machine. In this conventional example, as shown in FIG. 7, a slit light projector 50 that projects slit light toward a joint position portion of a joint, an imaging device 51 that captures an image of the projected slit light, and the imaging device 51 The output image signal is processed by the signal processing unit 52 to detect the seam position of the welded joint. The signal processing unit 52 includes a vertical synchronizing signal generating circuit 80, ay coordinate counter 81, a horizontal synchronizing signal generating circuit 82, an x coordinate counter 83, a pulse generating circuit 84, a comparing circuit 85, an averaging circuit 86, and a memory circuit 8.
7, a smoothing circuit 88, a subtraction circuit 89, and a darkest part detection circuit 90. In the signal processing unit 52,
By detecting the brightness of the image by the detected slit light,
The joint image included in the image signal of the image pickup device 51 is extracted using a threshold value based on the change of the brightness distribution in the slit light longitudinal direction of the image shown in FIG. 1, and the x-axis direction and the y-axis of the joint position are extracted from the extraction timing. The coordinate position of the direction is detected. Therefore, in the signal processing unit 52, as shown in the flowchart of FIG. 9, the slit light position and the slit light brightness are extracted, the brightness averaging process in the slit light width direction, the brightness distribution smoothing process in the slit light longitudinal direction, and the darkest position detection. Is sequentially executed to detect the seam position as the darkest part position P in FIG. Further, as shown in FIG. 10, at least three parallel linear lights a, b, and c are projected from the slit light projector 61 in the forward direction of the welding torch 60, and the linear lights a, b, and c are projected. The stepped state is imaged by the imaging device 62, the seam line position information to be welded at three positions is detected from the position 70, 71, 72 of the stepped portion, and the line equation f of the seam line is based on the seam line position information. (X, y, z) = 0 is calculated, the amount of deviation between the welding torch 60 and the seam line shown in FIG. 11 is detected, and the position of the welding torch 60 is controlled so that the amount of deviation is zero. Are known.

[0003]

Although it is possible to detect the seam line to be welded in both of the above-mentioned prior arts,
Since the darkest part of the image is recognized as the seam line,
If there is a temporary attachment part in the middle of the seam line, a problem occurs. That is, the surface of the tacking portion has irregularities over a wide range, and the irregularities cause variations in the dark portion of the image of the tacking portion, and it is possible to accurately determine the seam line by the conventional simple method. There is a problem that you cannot do it. In addition, it may be necessary to detect the position of the tacking portion, but it is not possible to meet such a request. In order to solve such a problem in the conventional technique, the present invention improves the method for detecting the seam position, and distinguishes the seam line from the temporary seam even when the temporary seam exists in the middle of the seam line. It is an object of the present invention to provide a seam detection method capable of estimating the position of a seam line.

[0004]

In order to achieve the above object, the present invention irradiates light from a light source to a seam position portion of a welded joint, photographs the seam position portion by an image pickup unit, and sends out from the image pickup unit. In the seam detection method for detecting the position of the seam line to be welded while discriminating the seam line of the joint and the tacked part by performing signal processing on the image signal that is Detecting each coordinate value of the seam position and calculating the straight line corresponding to the seam line to be welded in the image area based on each detected coordinate value sequence, and the straight line of each coordinate value sequence with respect to the calculated straight line A regression residual value is calculated, the calculated linear regression residual value is compared with a preset discrimination threshold value, and when the linear regression residual value is larger than the discrimination threshold value, it is detected as above. It is determined that each coordinate value sequence is a temporary part. And, when the linear regression residual value is less than determination threshold value is configured as a joint detection method in which each coordinate value sequence, which is the detected and judging that the above seam line. In this case, each coordinate value of the seam position may be obtained by setting the darkest position of the image signal in a plurality of sections set in the image area as the seam position. Further, a straight line corresponding to the seam line can be calculated by processing the coordinate values by the least square method. Furthermore, a method of calculating the linear regression residual value by calculating the residual sum of squares can be considered.

[0005]

According to the present invention, a straight line corresponding to the seam line to be welded in the image area is calculated based on each coordinate value sequence of the detected seam position, and the coordinate value sequence of each coordinate line for the calculated straight line is calculated. A linear regression residual value is calculated. Next, the calculated linear regression residual value is compared with a preset discrimination threshold value, and when the linear regression residual value is larger than the discrimination threshold value, the detected coordinate value sequences are When the linear regression residual value is smaller than the determination threshold, it is determined that each of the detected coordinate value sequences is the seam line. The seam position can be detected by the position of the darkest part of the image signal detected in a plurality of sections set in the image area. Further, a straight line corresponding to the seam line can be calculated by calculating the coordinate values by the least square method. Furthermore, it is also possible to calculate the linear regression residual value by calculating the residual sum of squares.

[0006]

Embodiments of the present invention will be described below with reference to the accompanying drawings for the understanding of the present invention. The following embodiments are examples of embodying the present invention and are not intended to limit the technical scope of the present invention. Here, FIG. 1 is a flowchart showing the entire seam detection method of the present invention, FIG. 2 is a block diagram showing a schematic configuration of a seam detection device 1 for carrying out the seam detection method of the present invention, and FIG. FIG. 4 is a flow chart showing a detection method, FIG. 4 is an explanatory view showing an image of a seam line imaged by the seam detection device 1 and an image of a temporary attachment portion, and FIG. 5 shows a plurality of processing regions set in the captured image. 6A and 6B are explanatory diagrams showing detection points (marked with X) detected in the plurality of processing regions. The seam detecting device 1 for detecting the seam of a butt joint having a zero root gap shown in FIG.
A light source 10 for irradiating diffused light such as a halogen light, and an image pickup device 11 are arranged on a seam line 22 to be welded where 21 is butted. The image signal of the imaging region obtained by the imaging device 11 is output to the signal processing unit 12.
By processing the image signal in the signal processing unit 12, the seam line 22 (FIG. 4A) and the tacking unit 23 in the imaging region are processed.
(FIG. 4 (B)) is determined. In the image obtained by the image pickup device 11, the seam line 22 is a straight line in the dark part, but since the surface of the tacking part 23 is uneven, the dark part in the image of the tacking part 23 varies in a scattered manner. In the present invention, the signal processing unit 12 discriminates between the seam line 22 and the temporary tacking unit 23 based on such a difference in image between the seam line 22 and the temporary tacking unit 23. The signal processing in the signal processing unit 12 will be described with reference to the flowcharts of FIGS. 1 and 3. First, in step S1 in FIG. 1, in the signal processing unit 12, as shown in FIG. 5, the obtained image is divided into a plurality of processing lines 25a, which are divided by a large number of dividing lines 24 ... b, c
… Set n. Next, the process proceeds to step S2, and the positions of the darkest part included in the image are detected at a plurality of positions for each of the processing areas 25a, b, c The darkest part position detection performed in step S2 is performed by the conventional Japanese Patent Publication No. 4-5924.
It is carried out in the same manner as the publication. That is, as shown in FIG. 3, after the image is captured in step S10, step S1
At 1, the image brightness is extracted. In step S12, the image brightness in the direction along the seam line 22 is averaged.
In step S13, the brightness distribution of the image in the direction intersecting the seam line 22 is smoothed. And these steps S
The position of the darkest part is detected from the difference between the calculation results of 12 and S13 (step S14).

In this way, each of the processing areas 25a, 25b, 25c
When the detection points 30 in the darkest part in the processing areas 25a, b, c ... n are obtained from the image signals in n (shown by x), the point sequence of each detection point 30 in the seam line 22 is It is distributed linearly as shown in FIG. On the other hand, in the image of the tacking portion 23, as shown in FIG. 6 (B), the unevenness of the surface of the tacking portion 23 causes the dark portion to scatter in a scattered manner, so the point sequences of the detection points 31 ... It will be distributed in the state.
Therefore, in this embodiment, paying attention to the difference in the distribution state between the detection points 30 of the seam line 22 and the detection points 31 of the tacking portion 23,
The seam line 22 and the tacking portion 23 are discriminated as follows. That is, in step S3 of FIG. 1, the least squares method is applied to the coordinate sequence of the plurality of detection points 30 and 31 detected in the image to calculate the line equation of the straight line corresponding to the seam line 22. Calculate the residual sum of squares of the linear regression for the straight line. The calculated residual sum of squares represents the degree of variation in the point sequence of the detection points 30 and 31 in the image.
Therefore, it is possible to discriminate between the seam line 22 and the tacking portion 23 by comparing the residual sum of squares required for linear regression with a preset discrimination threshold. Therefore, in step S4, if the residual sum of squares is larger than the discrimination threshold, it is recognized as the tacking portion 23, while if the residual sum of squares is smaller than the discrimination threshold, the seam to be welded. Line 22
Recognize that. When the tacking part 23 exists in the middle of the seam line 22, the seam line 22 can be estimated by extrapolation calculation using the seam line position series detected up to that point, so that the welding robot or automatic welding The welding torch of the machine can be guided along the seam line 22. In addition, since the tacking portion can be distinguished from the seam line, the tacking portion can be used for detecting an abnormality in the tacking position. It should be noted that the present invention is not limited to the above embodiment and can be modified in various ways. For example, although a diffused light source such as a halogen light is used as the light source 10 in the above embodiment, the present invention can be applied to the case where slit light is used as a light source. That is, a processing area is formed by slit light, and the darkest part in this slit light is detected. Also, the line equation of the seam line 22 can be obtained by not only dividing the image into a plurality of processing regions 25a, b, c ... N but also detecting a plurality of detection points 30 ... In the entire image.

[0008]

Since the seam detection method according to the present invention is configured as described above, when a temporary attachment portion is present in the middle of the seam line to be welded, the surface irregularity of the temporary attachment portion causes Since the linear regression residual value of the detection points detected by the temporary fitting part varies, the linear regression residual value of the detection points in the temporary fitting part becomes larger than the discrimination threshold, and it should be recognized as the temporary fitting part. You can As a result, by comparing the linear regression residual value and the discrimination threshold value, it is possible to discriminate between the seam line and the tack portion and detect the seam line.

[Brief description of drawings]

FIG. 1 is a flowchart showing an entire seam detection method according to an embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a seam detection device.

FIG. 3 is a flowchart showing a method for detecting the position of the darkest part.

4A is an explanatory view showing an image of a seam line imaged by a seam detection device, and FIG. 4B is an explanatory view showing an image of a temporary attachment portion.

FIG. 5 is an explanatory diagram showing a plurality of processing regions set in a captured image.

FIG. 6A is an explanatory diagram showing detection points of a seam line detected in a plurality of processing areas, and FIG. 6B is an explanatory diagram showing detection points of a temporary attachment portion detected in a plurality of processing areas. .

FIG. 7 is a block diagram showing a configuration of a first conventional example.

FIG. 8 is a graph showing the brightness distribution of an image by conventional slit light.

FIG. 9 is a flowchart showing a conventional method for detecting a seam line position.

FIG. 10 is a perspective view showing a configuration of a second conventional example.

FIG. 11 is an explanatory diagram showing a principle of calculating a line equation of a seam line to be welded.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Seam detection device 10 ... Light source 11 ... Imaging device 12 ... Signal processing part 22 ... Seam line 23 ... Temporary attachment part

Claims (4)

[Claims] 1. A joint joint line is obtained by irradiating a joint position portion of a welded joint with light from a light source, photographing the joint position portion by an image pickup unit, and processing the image signal sent from the image pickup unit. In the seam detection method for detecting the position of the seam line to be welded while distinguishing from the temporary part, each coordinate value of a plurality of seam positions is detected in the image region imaged by the imaging unit, and each detected A straight line corresponding to the seam line to be welded in the image area is calculated based on the coordinate value sequence, and
The linear regression residual value of each coordinate value sequence for the calculated straight line is calculated, and the calculated linear regression residual value is compared with a preset determination threshold value to determine the linear regression residual value. When it is larger than the threshold value, it is determined that the detected coordinate value sequence is a temporary part, and when the linear regression residual value is smaller than the determination threshold value, the detected coordinate value sequence is the seam value. A seam detection method characterized by determining that the line is a line. 2. The seam detection method according to claim 1, wherein each coordinate value of the seam position is obtained by setting the darkest position of the image signal in a plurality of sections set in the image area as a seam position. . 3. The seam detection method according to claim 1, wherein a straight line corresponding to the seam line is calculated by processing the coordinate values by a least square method. 4. The seam detection method according to claim 1, wherein the linear regression residual value is calculated by calculating a residual sum of squares.