Welding condition setting assistance device

Abstract

An image processing unit (141b) is provided to cause a computer (140) to execute the following steps: a sputtering candidate region detection step of detecting a sputtering candidate region for each of a plurality of input images of a workpiece (170) taken during arc welding, based on pixel values representing the brightness of pixels included in the input image; a reflected light region determination step of determining, based on color information of a prescribed reference pixel included in each sputtering candidate region, a reflected light region where reflected light of arc light is captured among the sputtering candidate regions detected in the sputtering candidate region detection step; and a sputtering count determination step of, for each input image, determining the number of sputtering candidate regions other than the reflected light region determined by the reflected light region determination step among the sputtering candidate regions detected in the sputtering candidate region detection step as the sputtering count.

Description

Technical Field

[0001] The present disclosure relates to a device for assisting in setting welding conditions when performing arc welding by generating an arc between a workpiece and an electrode by applying a voltage between the workpiece and the electrode. Background Art

[0002] The welding condition setting assistance device disclosed in Patent Document 1 detects spatter for each of a plurality of input images of a workpiece taken during arc welding, and determines the number of detected spatter based on pixel values representing the brightness of pixels included in each input image.

[0003] Prior Art Documents

[0004] Patent Documents

[0005] Patent Document 1: International Publication No. 2019 / 239644 Summary of the Invention

[0006] -Problems to be Solved by the Invention-

[0007] However, in the above Patent Document 1, since spatter is detected based on pixel values representing the brightness of pixels included in each input image, reflected light from peripheral devices of the arc light may be erroneously detected as spatter, and the number of spatter cannot be accurately determined. However, in Patent Document 1, no countermeasure is taken against this problem.

[0008] The present disclosure has been made in view of the above circumstances, and an object thereof is to be able to more accurately determine the number of spatter in an input image.

[0009] -Means for Solving the Problems-

[0010] In one aspect of the present disclosure, a welding condition setting assistance device is a device for assisting in setting welding conditions when performing arc welding by generating an arc between a workpiece and an electrode by applying a voltage between the workpiece and the electrode. The welding condition setting assistance device includes an image processing unit that executes the following steps: a spatter candidate region detection step of detecting a spatter candidate region for each of a plurality of input images of the workpiece taken during the arc welding, based on pixel values representing the brightness of pixels included in the input image; a reflected light region determination step of determining a reflected light region where reflected light of the arc light is captured, from the spatter candidate regions detected in the spatter candidate region detection step, based on color information of a prescribed reference pixel included in each spatter candidate region; and a spatter number determination step of determining the number of spatter in each input image as the number of spatter candidate regions in each input image detected in the spatter candidate region detection step, excluding the reflected light region determined in the reflected light region determination step.

[0011] According to this method, in the sputtering candidate area detection step, even if the area of the reflected light from the peripheral device where the arc light is captured is detected as the sputtering candidate area, the number of sputtering candidate areas other than the reflected light area where the reflected light of the arc light is captured among the sputtering candidate areas detected in the sputtering candidate area detection step is determined as the sputtering count. Therefore, the number of sputterings in the input image can be determined more accurately.

[0012] -Advantages of the Invention-

[0013] According to the welding condition setting assistance device according to the present disclosure, the number of sputterings in the input image can be determined more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a diagram showing a schematic configuration of a welding system including a computer that serves as a welding condition setting assistance device according to an embodiment of the present disclosure.

[0015] Figure 2 It is an explanatory diagram showing sputterings generated during arc welding.

[0016] Figure 3 It is a flowchart showing the steps of determining the sputtering count using a computer that serves as a welding condition setting assistance device according to an embodiment of the present disclosure.

[0017] Figure 4 It is an explanatory diagram exemplifying the detected sputterings.

[0018] Figure 5 It is a flowchart showing the steps of the first list correction process.

[0019] Figure 6 It is a flowchart showing the steps of the background bright point area removal process.

[0020] Figure 7 It is a flowchart showing the steps of the reflected light area removal process.

[0021] Figure 8 It is an explanatory diagram exemplifying the processed image. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Hereinafter, embodiments of the present disclosure will be described based on the drawings. The following description of the preferred embodiments is merely exemplary in nature and is not intended to limit the present disclosure, its applications, or its uses.

[0023] Figure 1Reference numeral 100 denotes a welding system. The welding system 100 includes a welding robot 110, a camera 120, a memory card 130 accommodated in the camera 120, a computer 140 that serves as a welding condition setting assistance device according to an embodiment of the present disclosure, and a card reader 150 connected to the computer 140.

[0024] As Figure 2 shown in the figure, the welding robot 110 includes a torch 111 that can hold a welding wire 160. By applying a voltage between a workpiece 170 held by a welding jig (clip) and the welding wire 160 serving as an electrode held by the torch 111, an arc A is generated between the workpiece 170 and the welding wire 160 to perform arc welding. During arc welding, the welded portion of the workpiece 170 melts and forms a molten pool 171, and spatter SP scatters from the molten pool 171. In addition, an ejection hole (not shown) for ejecting a shielding gas is provided at the tip of the torch 111.

[0025] The camera 120 is disposed at a position where it can photograph the entire spatter SP scattering region including the entire workpiece 170 via an ND (Neutral Density) filter (not shown), and saves the captured moving image in the memory card 130. In addition, the frame rate (shooting speed) of the camera 120 is set to 60 fps. Further, the focus, aperture, and shutter speed of the electronic shutter of the camera 120 are fixed.

[0026] The computer 140 includes a computer main body 141 and a display 142. The computer main body 141 includes a storage unit 141a and an image processing unit 141b.

[0027] The storage unit 141a of the computer main body 141 stores a learned model generated by supervised learning using a plurality of images of spatter SP and a plurality of images without spatter SP as teaching data. As a method of supervised learning for generating the learned model, for example, deep learning is used. In addition, the storage unit 141a also stores the moving image captured by the camera 120 and the still images obtained by segmenting the moving image.

[0028] The image processing unit 141b of the computer main body 141 reads the moving image saved in the memory card 130 inserted into the card reader 150 and saves it in the storage unit 141a. In addition, the image processing unit 141b divides the moving image saved in the storage unit 141a into still images (frames) and saves them as input images in the storage unit 141a. Further, the image processing unit 141b generates a processed image I (refer to Figure 8 ) obtained by subjecting each of the multiple input images saved in the storage unit 141a to a prescribed process. Details of the method for generating the processed image I will be described later.

[0029] The display 142 displays the processed image I generated by the image processing unit 141b of the computer main body 141.

[0030] Hereinafter, with reference to Figure 3 the steps for setting the welding conditions of the welding system 100 will be described.

[0031] First, in (S101), while the user is in a state where the welding robot 110 performs arc welding, the camera 120 is made to execute shooting, and the captured moving image is saved in the memory card 130. Thus, the moving image of the entire spatter SP scattering region including the entire workpiece 170 during arc welding is saved in the memory card 130. At this time, the frame rate of the captured and saved moving image is 60 fps.

[0032] Next, in (S102), the user removes the memory card 130 from the camera 120 and inserts it into the card reader 150, and causes the moving image saved in the memory card 130 to be transmitted from the card reader 150 to the computer main body 141. Then, the image processing unit 141b of the computer main body 141 receives the moving image transmitted from the card reader 150 and saves it in the storage unit 141a.

[0033] Next, in (S103), the user causes the image processing unit 141b of the computer main body 141 to divide the moving image during arc welding saved in the storage unit 141a into a plurality of still images (frames), and saves all the still images as input images in the storage unit 141a. At this time, for every 1 second of the moving image, 60 input images are saved in the storage unit 141a. Each input image is a color image, and the color of each pixel is represented by color information defined in the RGB color space.

[0034] Next, in (S104), the image processing unit 141b converts the input images saved in the storage unit 141a that have not yet been converted into grayscale images into grayscale images.

[0035] Next, in (S105), the image processing unit 141b performs a process of converting the pixel values of pixels with pixel values equal to or less than a specified threshold value Gs into values (0) representing black for the grayscale image obtained in (S104), and sets it as the spatter detection target image IM. Thus, pixels with pixel values representing brightness exceeding the specified threshold value Gs are determined as pixels other than black. And, for example Figure 4 as shown, the spatter candidate region SR of the high-brightness region HR including a plurality of consecutive pixels other than black in the spatter detection target image IM is detected (spatter candidate region detection step). Figure 4Among them, PB represents black pixels, and PW represents pixels other than black. The sputtering candidate region SR is a square region. The high-brightness region HR is located in a region other than the outer peripheral portion of the sputtering candidate region SR. In Figure 4 In the example of, the high-brightness region HR is located at the center of the sputtering candidate region SR. The detection of the sputtering candidate region SR here is performed using the learned model stored in the storage unit 141a. And, a first sputtering list ListGs for determining all the detected sputtering candidate regions SR is created.

[0036] Next, in (S106), the image processing unit 141b performs a process of converting the pixel values of pixels whose pixel values are equal to or less than a specified threshold Gm in the grayscale image obtained in (S104) into values (0) representing black. Then, the sputtering candidate regions SR including the high-brightness regions HR in which there are a plurality of consecutive pixels other than black in the processed image are detected. The detection of the sputtering candidate region SR here is also performed using the learned model stored in the storage unit 141a. In addition, the threshold Gm is set to a value larger than the threshold Gs. And, a second sputtering list ListGm for determining all the detected sputtering candidate regions SR is created.

[0037] Next, in (S107), the image processing unit 141b performs a process of converting the pixel values of pixels whose pixel values are equal to or less than a specified threshold Gl in the grayscale image obtained in (S104) into values (0) representing black. Then, the sputtering candidate regions SR including the high-brightness regions HR in which there are a plurality of consecutive pixels other than black in the processed image are detected. The detection of the sputtering candidate region SR here is also performed using the learned model stored in the storage unit 141a. In addition, the threshold Gl is set to a value larger than the threshold Gm. And, a third sputtering list ListGl for determining all the detected sputtering candidate regions SR is created. In addition, in the first to third sputtering lists ListGs, ListGm, ListGl, the sputtering candidate region SR is determined by the coordinates of the upper left pixel P, the horizontal length X, and the vertical length Y.

[0038] Here, since the larger the size of the sputtering SP, the greater its brightness, the sputtering SPs captured in the sputtering candidate regions SR determined in (S105) to (S107) are respectively sputtering SPs larger than the sizes corresponding to the thresholds Gs, Gm, Gl.

[0039] Next, in (S108), the image processing unit 141b creates a small sputtering list SmallS for determining the sputtering candidate regions SR obtained by removing the sputtering candidate regions SR determined in the second sputtering list ListGm from the sputtering candidate regions SR determined in the first sputtering list ListGs.

[0040] Next, in (S109), the image processing unit 141b determines a middle sputtering list MiddleS that determines a sputtering candidate region SR obtained by removing the sputtering candidate region SR determined in the third sputtering list ListGl from the sputtering candidate region SR determined in the second sputtering list ListGm.

[0041] Next, in (S110), the image processing unit 141b directly sets the third sputtering list ListGl as the large sputtering list LargeS.

[0042] Next, in (S111), the image processing unit 141b determines whether an input image (frame) for which conversion to a grayscale image has not been performed (processing in (S104) to (S110)) remains in the storage unit 141a. And, in the case where it remains, the process returns to (S104), while in the case where it does not remain, the process proceeds to (S112).

[0043] And, in (S112), the image processing unit 141b performs the following first list correction process.

[0044] Figure 5 Steps indicating the first list correction process performed in (S112).

[0045] In (S112), first, in (S201), the image processing unit 141b detects the center coordinates of all sputtering candidate regions SR detected in all input images stored in the storage unit 141a. In other words, the image processing unit 141b calculates the center coordinates of all sputtering candidate regions SR determined in the first sputtering list ListGs of all input images.

[0046] Next, in (S202), for each position (coordinate) on the input image, the image processing unit 141b calculates the number of times the center coordinates calculated in (S201) are the center coordinates of the sputtering candidate region SR as the sputtering detection count, and stores the distribution of the sputtering detection count.

[0047] Next, in (S203), the image processing unit 141b determines coordinates (positions) on the input image where the sputtering detection count calculated in (S202) is equal to or more than a specified reference count as background bright points. And, it determines a background bright point coordinate list that determines the coordinates of the background bright points.

[0048] Next, in (S204), the image processing unit 141b performs background bright spot region removal processing on all the input images stored in the storage unit 141a, which removes the sputtering candidate regions SR whose center coordinates are in the background bright spot coordinate list created in (S203) from the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS. In other words, the sputtering candidate regions SR removed by this background bright spot region removal processing are among all the sputtering candidate regions SR determined in the first sputtering list ListGs of all the input images, and the regions where the detection times of the sputtering candidate regions SR are above a specified reference number of times.

[0049] Here, with reference to Figure 6 the specific steps of the background bright spot region removal processing for the small sputtering list SmallS will be described. The background bright spot region removal processing for the medium sputtering list MiddleS and the large sputtering list LargeS is also performed through the same steps.

[0050] First, in (S301), the image processing unit 141b sets k = 0.

[0051] Next, in (S302), the image processing unit 141b sets k = k + 1.

[0052] Next, in (S303), the image processing unit 141b determines whether the center coordinates of the k-th sputtering candidate region SR in the small sputtering list SmallS exist in the background bright spot coordinate list. If the center coordinates of the k-th sputtering candidate region SR do not exist in the background bright spot coordinate list, it proceeds to (S305). If they exist, the k-th sputtering candidate region SR in the small sputtering list SmallS is determined as the background bright spot region and proceeds to (S304).

[0053] In (S304), the image processing unit 141b deletes the k-th sputtering candidate region SR from the small sputtering list SmallS.

[0054] In (S305), the image processing unit 141b determines whether the determination in (S303) has been performed for all the sputtering candidate regions SR in the small sputtering list SmallS. If not, it returns to (S302). If so, the background bright spot region removal processing ends.

[0055] Next, in (S113), the image processing unit 141b performs the following second list correction processing on all the input images. The second list correction processing performs the following reflected light region removal processing on the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS after the first list correction processing.

[0056] Here, with reference to Figure 7 the following, the specific steps of the reflected light region removal process for the small sputtering list SmallS will be described. The reflected light region removal processes for the medium sputtering list MiddleS and the large sputtering list LargeS are also carried out through the same steps.

[0057] First, in (S401), the image processing unit 141b calculates the center coordinates of all sputtering candidate regions SR determined in the small sputtering list SmallS stored in the storage unit 141a.

[0058] Next, in (S402), for all sputtering candidate regions SR determined in the small sputtering list SmallS, the image processing unit 141b calculates the average value of the red component, the average value of the green component, and the average value of the blue component of the color information of the specified reference pixels CP (refer to Figure 4 ) around the center coordinates calculated in (S401). Here, the reference pixels CP are all the pixels included in the square average calculation region AR that shares the center with the sputtering candidate region SR. The area of the average calculation region AR is set to 90% of the area of the sputtering candidate region SR.

[0059] Next, in (S403), the image processing unit 141b sets n = 0.

[0060] Next, in (S404), the image processing unit 141b sets n = n + 1.

[0061] And, in (S405), for the nth sputtering candidate region SR of the small sputtering list SmallS, with reference to the calculation results in (S402), the image processing unit 141b determines whether it satisfies the condition that the average value Gave of the green component of the color information of the reference pixels CP is greater than the product of the first set value α and the average value Rave of the red component of the color information of the reference pixels CP, and the average value Bave of the blue component of the color information of the reference pixels CP is greater than the product of the second set value β and the average value Rave of the red component of the color information of the reference pixels CP. The first set value α and the second set value β are set to values of 1 or more. Preferably, the first set value α is set to 1.35 or more, and the second set value β is set to 1.31 or more. And, when this condition is satisfied, the nth sputtering candidate region SR of the small sputtering list SmallS is determined as the reflected light region that captures the reflected light of the arc light, and proceeds to (S406). In other words, the sputtering candidate region SR that satisfies this condition is determined as the reflected light region. On the other hand, when this condition is not satisfied, it proceeds to (S407).

[0062] In (S406), the image processing unit 141b deletes the n-th sputtering candidate region SR from the small sputtering list SmallS.

[0063] In (S407), the image processing unit 141b determines whether the determination in (S405) has been performed for all the sputtering candidate regions SR in the small sputtering list SmallS. If not, it returns to (S404). If it has been performed, the reflected light region removal process ends.

[0064] Next, in (S114), the image processing unit 141b determines the number of sputtering candidate regions SR determined in the corrected small sputtering list SmallS as the sputtering count S of the smaller sputtering SP. Further, it determines the number of sputtering candidate regions SR determined in the corrected medium sputtering list SmallM as the sputtering count M of the medium-sized sputtering SP. Furthermore, it determines the number of sputtering candidate regions SR determined in the corrected large sputtering list LargeS as the sputtering count L of the larger sputtering SP. In this way, even if the light of the arc A is reflected from peripheral devices such as the welding jig (clip), the torch 111, and the main body of the welding robot 110, and the region capturing the reflected light from the peripheral device is detected as the sputtering candidate region SR in (S105) to (S107), the background bright point region determined in (S112) and the reflected light region determined in (S113) can be removed from the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS. Therefore, the sputtering counts S, M, and L can be determined more accurately.

[0065] In (S115), the image processing unit 141b determines the sum of the sputtering counts S, M, and L as the total sputtering count T. Further, as Figure 8 shown, the image processing unit 141b generates a processed image I obtained by performing the following respective processes on each input image. The respective processes are: a process of displaying the sputtering counts S, M, L, and the total sputtering count T at the upper left corner of the image; and a process of surrounding the sputtering candidate regions SR determined in the corrected small sputtering list SmallS in (S112) and (S113) with a blue rectangular frame F1, surrounding the sputtering candidate regions SR determined in the corrected medium sputtering list MiddleS in (S112) and (S113) with a yellow rectangular frame F2, and surrounding the sputtering SP determined in the corrected large sputtering list LargeS in (S112) and (S113) with a red rectangular frame F3. Additionally, Figure 8 in, the blue frame F1 is represented by a dotted line, the yellow frame F2 is represented by a solid line, and the red frame F3 is represented by a dashed line. Furthermore, Figure 8Among them, A’ represents the light of the arc A and the light of the arc A reflected from the flue gas. The total number of sputtering T is the number of the sputtering candidate regions SR determined from the first sputtering list ListGs after removing the background bright point regions determined for the three lists SmallS, MiddleS, and LargeS through (S112) and the reflected light regions determined for the three lists SmallS, MiddleS, and LargeS through (S113).

[0066] In this way, in (S112) to (S115), the image processing unit 141b performs the following steps: a reflected light region determination step of determining a reflected light region where the reflected light of the arc light is photographed based on the color information of the specified reference pixel CP included in each sputtering candidate region SR from the sputtering candidate region SR detected in (S105); a background bright point region determination step of determining a region where the number of times the sputtering candidate region SR detected in (S105) is equal to or more than a specified reference number as a background bright point region; and a sputtering number determination step of determining the number of the sputtering candidate regions SR after removing the reflected light region determined by the reflected light region determination step and the background bright point region determined by the background bright point region determination step among the sputtering candidate regions SR detected in (S105) as the total sputtering number T of each input image.

[0067] In (S116), the display 142 displays the processed image I generated in (S115). The user judges the suitability of the welding conditions such as the voltage value applied between the workpiece 170 and the bonding wire 160 by referring to the sputtering numbers S, M, L and the total sputtering number T displayed on the display 142. When the user determines that the welding conditions are inappropriate, the welding conditions are changed so as to reduce the number of the larger sputtering SP, and the processes of (S101) to (S116) are executed again. In addition, at this time on the display 142, the sputtering SP is displayed with frames F1, F2, F3 of colors corresponding to their sizes, so that the user can judge the reliability of the sputtering numbers S, M, L, T displayed on the display 142 by referring to these frames F1, F2, F3.

[0068] When the frame rate of the camera 120 is different from the frequency of the welding current, etc., the arc light may not necessarily be generated at the shooting timing of the input image. Therefore, although there are peripheral devices that reflect the arc light into the shooting range, the number of input images that capture the reflected light from these peripheral devices is small, and in (S112), it may not be possible to determine the area that captures the reflected light from these peripheral devices as the background bright point area. However, even in such a case, in the present embodiment, based on the color information of the reference pixel CP of the sputtering candidate area SR, the reflected light area in each input image is determined, and both the background bright point area and the reflected light area are removed from the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS. Therefore, compared with the case of only removing the background bright point area, the sputtering numbers S, M, and L can be determined more accurately.

[0069] In addition, when red external light from a constantly lit light source such as the light of a red LED (Light Emitting Diode) that constitutes an indicator of the machine or sunlight irradiates the peripheral devices existing in the shooting range, in (S113), it may not be possible to determine the area that captures the reflected light of such red external light as the reflected light area. However, even in such a case, in the present embodiment, the background bright point area where the detection count of the sputtering candidate area SR is equal to or more than a specified reference count is also determined, and both the background bright point area and the reflected light area are removed from the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS. Therefore, compared with the case of only removing the reflected light area, the sputtering numbers S, M, and L can be determined more accurately.

[0070] In addition, generally, the arc A is about 5000 to 20000 °C and emits pale light. On the other hand, the sputtering SP is about 700 to 1300 °C, which is lower than the arc A, and emits red light. Therefore, in the present embodiment, since the sputtering candidate area SR where the average value of the green component of the color information of the reference pixel CP is greater than the product of the first set value α and the average value of the red component, and the average value of the blue component of the color information of the reference pixel CP is greater than the product of the second set value β and the average value of the red component is determined as the reflected light area, it is easy to determine the sputtering candidate area SR that captures the reflected light of the arc A as the reflected light area rather than the sputtering SP.

[0071] In addition, since the image processing unit 141b determines the reflected light region by calculating the average value of the color information of the reference pixels CP included in the averaging region AR based on the average value of the regions other than the peripheral portion of the sputtering candidate region SR, the color information of the pixels in the darker region (the region that is not the high-brightness region HR) of the peripheral portion does not affect the determination of the reflected light region. Therefore, compared with the case of determining the reflected light region based on the average value of the color information of all the pixels in the sputtering candidate region SR, the color information of the pixels in the high-brightness region HR can be more significantly reflected in the determination of the reflected light region.

[0072] In addition, generally, even if a small sputtering SP adheres to the workpiece 170, the sputtering SP is easily removed by a metal brush or the like. On the other hand, if a large sputtering SP adheres to the workpiece 170, the sputtering SP cannot be removed unless it is ground by a grinding machine or the like, and the man-hours required for removal increase. In this embodiment, in (S116), since the numbers of sputtering SPs of multiple sizes, that is, the sputtering numbers S, M, L are displayed on the display 142, the user can refer to these sputtering numbers S, M, L and set the welding conditions so as to tolerate the occurrence of small sputtering SPs and suppress the occurrence of large sputtering SPs. Therefore, the man-hours required for removing the sputtering SPs can be reduced.

[0073] In addition, in (S105) to (S107), since the image processing unit 141b detects the sputtering candidate region SR by using a learned model generated by supervised learning with multiple images in which sputtering SPs are captured and multiple images in which sputtering SPs are not captured as teaching data, it is difficult to detect a region that is not a sputtering SP, such as a shielding gas or a part of the device, as the sputtering candidate region SR compared with the case of detecting all regions including a portion where multiple pixels other than black are continuous as the sputtering candidate region SR. Therefore, the possibility of false detection can be reduced.

[0074] In addition, generally, the larger the sputtering SP, the heavier it is and the lower its moving speed. Therefore, the larger the sputtering SP, the shorter the trajectory of the sputtering SP captured in one input image. Therefore, even when the photographing range is narrowed, a larger sputtering SP is more likely to be included in the photographing range than a smaller sputtering SP, and it is difficult to miss the detection of a large sputtering SP.

[0075] In addition, in this embodiment, in (S104), the image processing unit 141b temporarily converts the input image into a grayscale image, but the input image may not be converted into a grayscale image, and pixels whose pixel values representing brightness exceed a specified threshold may be directly determined from the input image.

[0076] In addition, in the present embodiment, in (S105) to (S107), the learned model is used to detect the sputtering candidate region SR. However, it is also possible to detect all regions with a specified shape that include regions with a continuous plurality of pixels other than black as the sputtering candidate region SR. In addition, it is also possible to detect all regions with a specified shape that include regions with a continuous number of pixels other than black greater than or equal to the first number and less than or equal to the second number (> the first number) as the sputtering candidate region SR. In such a case, regions including regions with a continuous number of pixels other than black exceeding the second number are not detected as the sputtering candidate region SR, so that it is possible to prevent regions that are not sputtering SP from being erroneously detected as the sputtering candidate region SR.

[0077] In addition, in the present embodiment, in (S201) to (S203), coordinates on the input image where the number of times corresponding to the center coordinates of the detected sputtering candidate region SR is equal to or greater than a specified reference number are determined as background bright points. However, it is also possible to determine coordinates on the input image where the number of times corresponding to a specified position other than the center of the detected sputtering candidate region SR is equal to or greater than a specified reference number as background bright points. For example, it is also possible to determine coordinates on the input image where the number of times corresponding to the upper left coordinates of the detected sputtering candidate region SR is equal to or greater than a specified reference number as background bright points, and perform a correction process for removing the sputtering candidate region SR with the upper left coordinates as background bright points for each list SmallS, MiddleS, and LargeS.

[0078] In addition, in the present embodiment, in (S201) to (S203), a background bright point coordinate list is created based on the first sputtering list ListGs of all input images, and this background bright point coordinate list is used for the background bright point region removal process of the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS in (S204). However, it is also possible to use the background bright point coordinate list created based on the small sputtering list SmallS of all input images for the background bright point region removal process of the small sputtering list SmallS, use the background bright point coordinate list created based on the medium sputtering list MiddleS of all input images for the background bright point region removal process of the medium sputtering list MiddleS, and use the background bright point coordinate list created based on the large sputtering list LargeS of all input images for the background bright point region removal process of the large sputtering list LargeS.

[0079] In addition, in the present embodiment, the average value calculation region AR is set to a square region sharing the center with the sputtering candidate region SR. However, for regions other than the peripheral portion in the sputtering candidate region SR, it may be set to other regions such as a circular region sharing the center with the sputtering candidate region SR. In addition, the area of the average value calculation region AR is set to 90% of the area of the sputtering candidate region SR. However, if it is less than 100%, it may be set to other ratios.

[0080] In addition, in the present embodiment, the image processing unit 141b performs the first and second list correction processes on the small sputtering list SmallS, the medium sputtering list MiddleS, and the large sputtering list LargeS in (S112) and (S113). However, it may only perform the second list correction process in (S113) without performing the first list correction process in (S112).

[0081] In addition, in the present embodiment, for the input image, the detected sputtering SP is processed with frames F1, F2, and F3 attached thereto to obtain a processed image I. However, it may also be processed with marks other than frames F1, F2, and F3 attached thereto to obtain a processed image I.

[0082] In addition, in the present embodiment, the image processing unit 141b of the computer main body 141 receives a moving image including the input image from the card reader 150. However, it may also receive it from other information communication devices.

[0083] In addition, in the present embodiment, the present disclosure is applied to arc welding using the welding robot 110. However, the present disclosure can also be applied to the case where the torch is operated manually.

[0084] In addition, the present disclosure can also be applied to either the case where the voltage applied between the workpiece 170 and the bonding wire 160 is not a pulse voltage or the case where it is a pulse voltage.

[0085] Industrial Applicability

[0086] The welding condition setting assistance device of the present disclosure can more accurately determine the number of sputterings in the input image and is useful as a device for assisting in setting welding conditions when generating an arc between a workpiece and an electrode by applying a voltage between the workpiece and the electrode to perform arc welding.

[0087] -Symbol Explanation-

[0088] 140 Computer (Welding Condition Setting Assistance Device)

[0089] 141b Image Processing Unit

[0090] 160 Bonding Wire (Electrode)

[0091] 170 workpieces

[0092] A arc

[0093] SP sputtering

[0094] SR sputtering candidate area

[0095] HR high-brightness area

[0096] CP reference pixel

[0097] T total number of sputterings

[0098] α first set value

[0099] β second set value.

Claims

1. A welding condition setting assistance device, which assists in setting welding conditions when arc welding is performed by applying a voltage between a workpiece and an electrode to generate an arc between the workpiece and the electrode. The welding condition setting assistance device includes an image processing unit, and the image processing unit performs the following steps: A sputtering candidate region detection step, for each of a plurality of input images of the workpiece taken during the arc welding, based on pixel values representing the brightness of pixels included in the input image, detects sputtering candidate regions; A reflected light region determination step, based on the color information of a prescribed reference pixel included in each sputtering candidate region, determines, from the sputtering candidate regions detected in the sputtering candidate region detection step, a reflected light region where the reflected light of the arc light is photographed; And A sputtering count determination step, determines, as the sputtering count of each input image, the number of sputtering candidate regions of each input image detected in the sputtering candidate region detection step, excluding the reflected light region determined in the reflected light region determination step.

2. The welding condition setting assistance device according to claim 1, wherein the color information is defined by an RGB color space, the reflected light region determination step determines, as the reflected light region, a sputtering candidate region that satisfies the following condition: the average value of the green component of the color information of the reference pixel is greater than the product of a first set value greater than 1 and the average value of the red component of the color information of the reference pixel, and the average value of the blue component of the color information of the reference pixel is greater than the product of a second set value greater than 1 and the average value of the red component of the color information of the reference pixel.

3. The welding condition setting assistance device according to claim 1 or 2, wherein the sputtering candidate region includes: a high brightness region where a plurality of consecutive pixels having pixel values representing brightness exceeding a prescribed threshold value are included.

4. The welding condition setting assistance device according to claim 3, wherein the high brightness region is located in a region other than the peripheral portion of the sputtering candidate region, the prescribed reference pixel is a pixel in a region other than the peripheral portion of each sputtering candidate region.

5. The welding condition setting assistance device according to claim 1 or 2, wherein the image processing unit further performs: a background bright point region determination step, determines, as a background bright point region, a region where the number of detections as the sputtering candidate region in the sputtering candidate region detection step is equal to or more than a prescribed reference number, the sputtering count determination step determines, as the sputtering count of each input image, the number of sputtering candidate regions of each input image detected in the sputtering candidate region detection step, excluding the reflected light region determined in the reflected light region determination step and the background bright point region determined in the background bright point region determination step.

Images