JP2020142265A - Welding condition setting support apparatus - Google Patents

Abstract

To simplify processing for identifying the number of sputter generated in arc welding, and reduce equipment cost.SOLUTION: A computer 140 is provided with an image processing unit 141b generating a masking-processed image obtained by drawing a pixel value of a mask image from a pixel value of each pixel of an input image photographing a workpiece 170 in arc welding, converting the masking-processed image into a gray scale image, giving the processing for converting the pixel value of the pixel where the pixel value exceeds a predetermined threshold into a value showing the black to the gray scale image, detecting sputter being made of consecutive multiple pixels from the pixel other than the black among images after the processing, and identifying the number of sputter detected.SELECTED DRAWING: Figure 1

Description

The present invention relates to a device that assists in setting welding conditions when performing arc welding by generating an arc between the work and the electrode by applying a voltage between the work and the electrode.

In Patent Document 1, an arc is photographed with a video camera at the time of arc welding to acquire a plurality of input images, the brightness distribution on a predetermined detection line in the image is captured for each input image, and the detection line is on the detection line based on the brightness distribution. A technique for identifying the number of spatters generated during arc welding by detecting spatters that have passed through is disclosed.

Japanese Unexamined Patent Publication No. 2009-28775

However, in Patent Document 1, since spatter that has passed on the detection line is detected based on a plurality of input images, if the shooting speed of the video camera is too slow, the accuracy of the number of spatters specified deteriorates. Therefore, a video camera having a high shooting speed is required, and the equipment cost increases.

The present invention has been made in view of this point, and an object of the present invention is to reduce equipment costs.

In one aspect of the present invention, the device that assists in setting welding conditions when performing arc welding by generating an arc between the work and the electrode by applying a voltage between the work and the electrode is the arc. Spatter detection processing that detects spatter composed of a plurality of continuous pixels whose brightness pixel values exceed a predetermined threshold value for a region excluding a predetermined mask region in an input image obtained by photographing the work during welding. And an image processing unit that executes a spatter number specifying process for specifying the number of spatters detected by the spatter detection process.

According to this aspect, since sputtering can be detected using only one input image, a video camera having a high shooting speed becomes unnecessary, and equipment cost can be reduced.

Further, since the number of spatters larger than the predetermined size corresponding to the threshold value is specified, the welding conditions are set so that the number of spatters larger than the predetermined size is reduced by referring to the specified number of spatters. Easy to set. Therefore, it becomes easy to reduce spatter larger than a predetermined size.

In addition, by setting the area where the arc light and the welding jig are reflected in the input image as the mask area, it is possible to prevent false detection of non-spattered objects as sputtering, so that the number of sputtering can be specified more accurately. It will be possible.

According to the welding condition setting support device according to the present invention, spatter can be detected by using only one input image, so that a video camera having a high shooting speed becomes unnecessary and equipment cost can be reduced.

Further, since the number of spatters larger than the predetermined size corresponding to the threshold value is specified, the welding conditions are set so that the number of spatters larger than the predetermined size is reduced by referring to the specified number of spatters. Easy to set. Therefore, it becomes easy to reduce spatter larger than a predetermined size.

In addition, by setting the area where the arc light and the welding jig are reflected in the input image as the mask area, it is possible to prevent false detection of non-spattered objects as sputtering, so that the number of sputtering can be specified more accurately. It will be possible.

It is a figure which shows the schematic structure of the welding system provided with the computer as the welding condition setting support device which concerns on embodiment of this invention. It is explanatory drawing which shows the spatter generated at the time of arc welding. It is a flowchart which shows the procedure of setting the welding condition using the computer as the welding condition setting support device which concerns on embodiment of this invention. It is a flowchart which shows the procedure of a mask processing. It is explanatory drawing which illustrates the grayscale image of the superimposition image. It is explanatory drawing which illustrates the binarized image of the grayscale image of the superposed image. It is explanatory drawing which shows the method of expanding the white part of a binarized image. It is explanatory drawing which illustrates the mask image. It is explanatory drawing which illustrates the detected sputtering. It is explanatory drawing which illustrates the processed image.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. The following description of preferred embodiments is merely exemplary and is not intended to limit the present invention, its applications or its uses.

FIG. 1 shows a welding system 100. The welding system 100 includes a welding robot 110, a video camera 120, a memory card 130 housed in the video camera 120, a computer 140 as a welding condition setting support device according to an embodiment of the present invention, and the computer 140. It is equipped with a card reader 150 connected to.

As shown in FIG. 2, the welding robot 110 includes a welding torch 111 capable of holding the welding wire 160, and welds the work 170 held by the welding jig (clamp) and the welding as an electrode held by the welding torch 111. By applying a voltage between the wire 160 and the work 170, an arc A is generated between the work 170 and the welding wire 160 to perform arc welding. At the time of arc welding, the portion to be welded of the work 170 is melted to form a molten pool 171, and spatter SP is scattered from the molten pool 171. The tip of the welding torch 111 is provided with an ejection hole (not shown) for ejecting the shield gas.

The video camera 120 is installed at a position where the entire scattering region of the sputtering SP including the entire work 170 can be photographed via an ND (Neutral Density) filter (not shown), and the captured moving image is stored in the memory card 130. The frame rate (shooting speed) of the video camera 120 is set to 60 fps. Further, the focus, aperture, and shutter speed of the electronic shutter of the video camera 120 are fixed.

The computer 140 includes a computer main body 141 and a display 142 as a display unit. The computer main body 141 includes a storage unit 141a and an image processing unit 141b.

The storage unit 141a of the computer main body 141 stores a trained model generated by supervised learning using a plurality of images showing the sputter SP and a plurality of images not showing the spatter SP as supervised data. .. As a supervised learning method for generating a trained model, for example, deep learning is used. The storage unit 141a further stores the moving image taken by the video camera 120 and the still image obtained by dividing the moving image.

The image processing unit 141b of the computer main body 141 reads out the moving image stored in the memory card 130 inserted in the card reader 150 and stores it in the storage unit 141a. Further, the image processing unit 141b divides the moving image stored in the storage unit 141a into still images (frames) and stores the moving image as an input image in the storage unit 141a. Further, the image processing unit 141b determines the mask area based on the plurality of input images stored in the storage unit 141a, detects the spatter SP from the area excluding the mask area in the input image, and detects the number of sputter SPs. (Mask area determination process, spatter detection process, spatter number identification process). The sputter SP is detected using the trained model stored in the storage unit 141a. In addition, the image processing unit 141b generates a processed image I (see FIG. 10) in which a predetermined process is applied to the input image (image generation process). Details of the method for determining the mask region, the method for detecting the sputtering SP, the method for specifying the number of sputtering SPs, and the method for generating the processed image I will be described later.

The display 142 displays the processed image I generated by the image processing unit 141b of the computer main body 141.

Hereinafter, the procedure for setting the welding conditions of the welding system 100 will be described with reference to FIG.

First, in (S101), the user causes the video camera 120 to perform shooting with the welding robot 110 performing arc welding, and saves the shot moving image in the memory card 130. As a result, the moving image of the entire scattering region of the sputtering SP including the entire work 170 at the time of arc welding is stored in the memory card 130.

Next, in (S102), the user takes out the memory card 130 from the video camera 120, inserts it into the card reader 150, and causes the card reader 150 to transfer the moving image stored in the memory card 130 to the computer main body 141. Then, the image processing unit 141b of the computer main body 141 receives the moving image transferred from the card reader 150 and stores it in the storage unit 141a.

Next, in (S103), the user causes the image processing unit 141b of the computer main body 141 to divide the moving image at the time of arc welding stored in the storage unit 141a into a plurality of still images (frames), and inputs the still image. Is stored in the storage unit 141a.

Then, in (S104), the image processing unit 141b performs the following mask processing.

FIG. 4 shows the procedure of the mask processing performed in (S104).

In (S104), first, in (S201), the image processing unit 141b converts all the input images (frames) stored in the storage unit 141a into grayscale images. The pixel value of the grayscale image indicates the brightness in 8 bits, the pixel value indicating black is 0, and the pixel value indicating white is 255.

Next, in (S202), the image processing unit 141b generates a superposed image based on the grayscale images of all the input images obtained in (S201). The pixel value of each pixel of the superimposed image is the average of the pixel values of all the grayscale images obtained in (S201). FIG. 5 illustrates the superimposed image obtained in (S202).

Next, in (S203), the image processing unit 141b converts the superimposed image generated in (S202) into a binarized image. Specifically, the pixel value of a pixel whose pixel value of the superimposed image is equal to or less than a predetermined reference value is converted into a value (0) indicating black, and the pixel value of a pixel whose pixel value of the superimposed image exceeds a predetermined reference value is converted. , Converted to a value indicating white (255). FIG. 6 illustrates the binarized image obtained in (S203).

In (S204), the image processing unit 141b generates a mask image in which the white portion of the binarized image obtained in (S203) is expanded. Specifically, as shown in FIG. 7, a process of converting the black pixels included in the 7 × 7 region RE centered on the white pixels WP of the binarized image obtained in (S203) into white pixels is performed. FIG. 8 illustrates the mask image obtained in (S204). As a result, the region of the white pixel in the mask image is determined as the mask region.

Then, in (S205), the image processing unit 141b performs a process of subtracting the pixel value of each pixel of the mask image from the pixel value of each pixel on each input image stored in the storage unit 141a, and the processed image. Is stored in the storage unit 141a as an image after masking. In the image after mask processing, the mask area determined in (S204) becomes black pixels. As a result, the mask processing is completed.

After the mask processing is completed, in (S105), among the images after mask processing stored in the storage unit 141a, the image processing unit 141b has not yet performed the processing of (S105) to (S111). To a grayscale image.

Next, in (S106), the image processing unit 141b sets the pixel value of the pixel whose pixel value is equal to or less than the predetermined threshold value Gs to the value (0) indicating black with respect to the grayscale image obtained in (S105). A process of conversion is performed to obtain an image IM to be spatter-detected. As a result, pixels whose brightness-indicating pixel values exceed a predetermined threshold value Gs are specified as pixels other than black (specific pixels). Then, for example, as shown in FIG. 9, a region composed of a plurality of continuous pixels is detected as a sputtering SP from pixels (specific pixels) other than black included in the sputtering detection target image IM. In FIG. 9, PB indicates a black pixel and PW indicates a non-black pixel. The detection of the sputtering SP here is performed using the trained model stored in the storage unit 141a. Then, the first spatter list ListGs for specifying all the detected spatter SPs is created.

Next, in (S107), the image processing unit 141b sets the pixel value of the pixel whose pixel value is equal to or less than a predetermined threshold Gm (first threshold) to black with respect to the grayscale image obtained in (S105). A process of converting to the indicated value (0) is performed, and in the processed image, a region composed of a plurality of consecutive pixels is detected as a spatter SP from pixels other than black (specific pixels). The detection of the sputtering SP here is also performed using the trained model stored in the storage unit 141a. The threshold value Gm is set to a value larger than the threshold value Gs. Then, a second sputter list ListGm for specifying all the detected spatter SPs is created.

Next, in (S108), the image processing unit 141b sets the pixel value of the pixel whose pixel value is equal to or less than a predetermined threshold Gl (second threshold) to the grayscale image obtained in (S105) to black. A process of converting to the indicated value (0) is performed, and in the processed image, a region composed of a plurality of consecutive pixels is detected as a spatter SP from pixels other than black (specific pixels). The detection of the sputtering SP here is also performed using the trained model stored in the storage unit 141a. The threshold value Gl is set to a value larger than the threshold value Gm. Then, a third sputter list ListGl for specifying all the detected spatter SPs is created. In the first to third sputtering lists ListGs, ListGm, and ListGl, the sputtering SP is specified by the coordinates.

Here, since the brightness of the sputtering SP increases as its size increases, the sputtering SPs specified in (S106) to (S108) are larger than the sizes corresponding to the threshold values Gs, Gm, and Gl, respectively. It is the number of spatter SPs.

Further, in the image after mask processing, the mask region determined in (S204) is a black pixel, so that the detection of the sputtering SP in (S106) to (S108) is a region in which the mask region is omitted from the input image. Will be done for.

Next, in (S109), the image processing unit 141b specifies a sputter SP excluding the sputter SP specified by the second sputter list ListGm from the sputter SP specified by the first sputter list ListGs. Create the list SmallS. Further, the image processing unit 141b specifies the number of sputters SP specified by the small spatter list SmallS as the number of spatters S of the small spatter SP.

Subsequently, in (S110), the image processing unit 141b specifies a sputter SP excluding the sputter SP specified by the third sputter list ListGl from the sputter SP specified by the second sputter list ListGm. Create a list MiddleS. Further, the image processing unit 141b specifies the number of sputters SP specified by the medium spatter list Middle S as the number of spatters M of the sputter SP having a medium size.

Subsequently, in (S111), the image processing unit 141b uses the third sputtering list ListGl as a large sputtering list LargeS as it is, and sets the number of sputtering SPs specified by the large sputtering list LargeS to the large sputtering SP. It is specified as the number of spatters L.

Next, in (S112), the image processing unit 141b determines whether or not the masked image (frame) that has not yet been processed in (S105) to (S111) remains in the storage unit 141a. Then, if it remains, the process returns to (S105), while if it does not remain, the process proceeds to (S113).

In (S113), the image processing unit 141b specifies the total of the sputtering numbers S, M, and L as the total sputtering number T. Further, the image processing unit 141b performs a process of imparting a frame number and a total sputtering number T to the upper left corner of the image for each input image, and a sputtering SP specified by the small sputtering list SmallS as a blue rectangle. Processing in which the spatter SP specified by the medium spatter list MiddleS is surrounded by the yellow rectangular frame F2, and the sputter SP specified by the large sputter list LargeS is surrounded by the red rectangular frame F3. To generate the processed image I illustrated in FIG. In FIG. 10, the blue frame F1 is indicated by a double line, the yellow frame F2 is indicated by a solid line, and the red frame F3 is indicated by a broken line. Further, in FIG. 10, A'indicates the region in which the light of the arc A itself and the work 170, the welding jig, and the fume illuminated by the light of the arc A are reflected. Since the region A'is made into a mask region by the masking process, it is possible to prevent erroneous detection of objects other than the sputtering SP as the sputtering SP, and it is possible to more accurately specify the number of the sputtering SP. Further, in (S201) to (S204), since the mask area is determined based on a plurality of input images of the work 170 taken during arc welding, objects such as the work 170 and the welding jig that are easily erroneously detected as the spatter SP Even when the relative position with respect to the video camera 120 is not fixed, it is possible to more reliably suppress the false detection of an object other than the spatter SP as the spatter SP.

In (S114), the display 142 displays the processed image I generated in (S113). The user determines the suitability of welding conditions such as the voltage value applied between the work 170 and the welding wire 160 by referring to the total sputtering number T displayed on the display 142. When the user determines that the welding conditions are inappropriate, the user changes the welding conditions so as to reduce the number of large sputtering SPs, and executes the processes (S101) to (S114) again. Further, at this time, since the spatter SP is displayed on the display 142 with the frames F1, F2, F3 of colors corresponding to the sizes thereof, the user can refer to these frames F1, F2, F3 to display. The reliability of the spatter number T displayed on 142 can be determined.

Therefore, according to the present embodiment, in (S105) to (S111), the image processing unit 141b specifies the number of spatter SPs using only one input image (image after mask processing), and therefore, Patent Document 1 As compared with the case where the number of spatter SPs is specified by using a plurality of input images as described above, the process can be simplified. Further, the number of sputtering SPs can be specified by using only one input image, and the video camera 120 does not require a video camera having a high shooting speed, so that the equipment cost can be reduced.

Further, in general, even if a small spatter SP adheres to the work 170, the spatter SP can be easily removed with a metal brush or the like, while if a large spatter SP adheres to the work 170, the spatter SP can be removed without polishing with a grinder or the like. However, the man-hours required for removal increase.

According to the present embodiment, in (S106) to (S108), the image processing unit 141b detects the sputtering SP with a plurality of images in which the sputtering SP is shown and a plurality of images in which the sputtering SP is not shown. Since this is performed using a trained model generated by supervised learning using the above as supervised data, compared to the case where all the parts where multiple continuous pixels other than black are continuous are detected as sputtering SP, the shield gas and the device It is difficult to detect a part that is not a sputter SP as a sputter SP. Therefore, the possibility of erroneous detection can be reduced.

Further, in general, the larger the sputtering SP, the heavier the sputter SP, and the lower the moving speed. Therefore, the trajectory of the sputtering SP captured in one input image becomes shorter as the sputtering SP increases. Therefore, even when the imaging range is narrowed, the large sputtering SP is more likely to be contained within the imaging range than the small sputtering SP, and detection omission of the large sputtering SP is unlikely to occur.

In this embodiment, the image processing unit 141b once converted the masked image into a grayscale image in (S105), but did not convert the masked image into a grayscale image, but directly from the masked image. A pixel whose brightness-indicating pixel value exceeds a predetermined threshold may be specified as a specific pixel.

Further, in the present embodiment, in (S106) to (S108), the sputter SP is detected using the trained model, but all the regions where a plurality of pixels other than black are continuous are detected as the sputter SP. You may. Further, the entire region in which pixels other than black are continuous by the number of the first number or more and the number of the second number (> the first number) or less may be detected as the sputtering SP. In this case, since the region in which pixels other than black exceed the second number and are continuous is not detected as the sputtering SP, it is possible to prevent erroneous detection of a non-sputtering SP as the sputtering SP.

Further, in the present embodiment, the input image is processed to attach the frames F1, F2, F3 to the detected sputter SP to obtain the processed image I, but the input image is processed to add a mark other than the frames F1, F2, F3. The processed image I may be used.

Further, in the present embodiment, although the image processing unit 141b of the computer main body 141 receives the moving image including the input image from the card reader 150, it may be received from another information transmission device.

Further, in the present embodiment, the present invention is applied to arc welding using the welding robot 110, but the present invention can also be applied to the case where the welding torch is manually operated.

Further, in the present embodiment, the present invention is applied when the voltage applied between the work 170 and the welding wire 160 is not a pulse voltage, but the present invention is applied between the work 170 and the welding wire 160. It can also be applied when the voltage is a pulse voltage.

Further, in the present embodiment, in (S105) to (S108), the sputter SP was detected using the image after the mask treatment. However, once the sputtering SP is detected from the input image and the sputtering SP located in the mask region is omitted from the list of detected sputtering SPs, the sputtering SP is detected in the region where the mask region is omitted in the input image. You may.

Further, in the present embodiment, the input image is processed to give only the display of the frame number and the total sputtering number T to obtain the processed image I. However, in addition to the frame number and the total sputtering number T, the sputtering number S, The processed image I may be obtained by performing processing to give the display of M and L.

The welding condition setting support device of the present invention can reduce the equipment cost, facilitate the reduction of spatters larger than a predetermined size, and prevent erroneous detection of non-spatters as spatters. It is useful as a device that assists in setting welding conditions when performing arc welding by generating an arc between the work and the electrode by applying a voltage between the workpiece and the electrode.

140 Computer (welding condition setting support device)
141a Storage unit
141b Image processing unit 142 Display (display unit)
160 Welding wire (electrode)
170 Work A Arc SP Sputter I Processed image F1, F2, F3 Frame (mark)

Claims (6)

It is a device that supports the setting of welding conditions when performing arc welding by generating an arc between the work and the electrode by applying a voltage between the work and the electrode.
Spatter that detects spatter composed of a plurality of continuous pixels whose brightness pixel values exceed a predetermined threshold value in a region excluding a predetermined mask region in an input image obtained by photographing the work during arc welding. A welding condition setting support device including an image processing unit that executes a detection process and a spatter number specifying process for specifying the number of spatters detected by the spatter detection process. In the welding condition setting support device according to claim 1,
The image processing unit is a welding condition setting support device that further executes a mask region determination process for determining the mask region based on a plurality of input images obtained by photographing the work during the arc welding. In the welding condition setting support device according to claim 1 or 2.
The image processing unit sets the predetermined threshold value to a first threshold value and a second threshold value larger than the first threshold value, and executes the sputtering detection process.
The sputtering number specifying process is detected by a sputtering detection process in which the predetermined threshold value is set to the second threshold value among the sputterings detected by the sputtering detection process in which the predetermined threshold value is set to the first threshold value. A welding condition setting support device for specifying the number of spatters excluding the spatters and the number of spatters detected by the spatter detection process in which the predetermined threshold value is set to the second threshold value. .. In the welding condition setting support device according to any one of claims 1 to 3.
The spatter detection process by the image processing unit uses a trained model generated by supervised learning using a plurality of images showing spatter and a plurality of images without spatter as teacher data. A welding condition setting support device characterized in that it detects. In the welding condition setting support device according to any one of claims 1 to 4.
A welding condition setting support device further provided with a display unit for displaying the number of spatters specified by the spatter number specifying process of the image processing unit. In the welding condition setting support device according to claim 5.
The image processing unit further executes an image generation process for generating a processed image in which the input image is processed to mark the sputtering detected by the sputtering detection process.
The display unit is a welding condition setting support device characterized by displaying a processed image generated by the image generation process.