Defect detection method, arc welding method, defect detection device, arc welding device, program, and recording medium

Abstract

The present invention provides a defect detection method, an arc welding method, a defect detection device, an arc welding device, a program, and a recording medium that can detect defects with high accuracy. [Solution] The defect detection method according to the embodiment includes an acquisition step, a detection step, and a determination step. In the acquisition step, an image of the area around the molten pool during arc welding is acquired. In the detection step, a presumed defect area and the molten pool, which are presumed to be defects, are detected in the image. In the determination step, it is determined whether or not the presumed defect area is a defect based on the positional relationship between the presumed defect area and the molten pool. In the determination step, if the presumed defect area is at a predetermined position relative to the molten pool, it is determined that the presumed defect area is a defect, and if the presumed defect area is not at the predetermined position relative to the molten pool, it is determined that the presumed defect area is not a defect.

Description

Technical Field

[0001] Embodiments of the present invention relate to a defect detection method, an arc welding method, a defect detection device, an arc welding device, a program, and a recording medium.

Background Art

[0002] For example, when joining a heat sink and fins by arc welding, there may be a hole defect in the bead after welding. As a means for suppressing the occurrence of the hole defect, it is conceivable to detect and repair the defect generated in the molten pool during welding. However, when detecting a defect, over-detection may occur due to arc light or the like. There is a need for a defect detection method capable of accurately detecting defects.

Prior Art Documents

Non-Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The problem to be solved by the present invention is to provide a defect detection method, an arc welding method, a defect detection device, an arc welding device, a program, and a recording medium capable of accurately detecting defects.

Means for Solving the Problems

[0005] The defect detection method according to the embodiment includes an acquisition step, a detection step, and a determination step. In the acquisition step, an image of the area around the molten pool during arc welding is acquired. In the detection step, a presumed defect area and the molten pool, which are presumed to be defects, are detected in the image. In the determination step, it is determined whether or not the presumed defect area is a defect based on the positional relationship between the presumed defect area and the molten pool. In the determination step, if the presumed defect area is at a predetermined position relative to the molten pool, it is determined that the presumed defect area is a defect, and if the presumed defect area is not at the predetermined position relative to the molten pool, it is determined that the presumed defect area is not a defect. [Brief explanation of the drawing]

[0006] [Figure 1] This is a schematic perspective view showing a defect detection device and an arc welding device according to an embodiment. [Figure 2] This is a schematic block diagram showing a defect detection device and an arc welding device according to the embodiment. [Figure 3] This is a schematic side view showing the position of the camera in the defect detection device and arc welding device according to the embodiment. [Figure 4] This is a schematic plan view showing the position of the camera in the defect detection device and arc welding device according to the embodiment. [Figure 5] This is an explanatory diagram schematically showing the first image acquired by the first camera of the defect detection device and arc welding device according to the embodiment. [Figure 6] This is an explanatory diagram schematically showing a second image acquired by a second camera of a defect detection device and an arc welding device according to an embodiment. [Figure 7] This is a flowchart illustrating an example of a defect detection method according to the first embodiment. [Figure 8] This is a flowchart illustrating an example of a defect detection method according to the second embodiment. [Figure 9] This is a flowchart illustrating an example of a defect detection method according to the third embodiment. [Figure 10]This is a flowchart illustrating an example of a defect detection method according to the fourth embodiment. [Figure 11] This is a flowchart illustrating an example of a defect detection method according to the fifth embodiment. [Figure 12] This is a flowchart showing an example of the first determination process of the defect detection method according to the fifth embodiment. [Figure 13] This is a flowchart showing an example of the second determination process of the defect detection method according to the fifth embodiment. [Figure 14] This is a flowchart illustrating an example of an arc welding method according to the embodiment. [Modes for carrying out the invention]

[0007] Each embodiment of the present invention will be described below with reference to the drawings. Drawings are schematic or conceptual, and the relationships between the thickness and width of each part, as well as the ratios of the sizes of different parts, are not necessarily identical to those of reality. Even when representing the same part, the dimensions and ratios may differ between drawings. In this specification and in each figure, elements similar to those described above are denoted by the same reference numerals with respect to previously shown figures, and detailed explanations are omitted as appropriate.

[0008] Furthermore, in order to make the explanation easier to understand, the arrangement and configuration of each part will be described using the XYZ Cartesian coordinate system. The X, Y, and Z axes are mutually orthogonal. The direction in which the X axis extends will be called the "X direction," the direction in which the Y axis extends will be called the "Y direction," and the direction in which the Z axis extends will be called the "Z direction." Also, in order to make the explanation easier to understand, the direction of the arrow in the Z direction will be considered upward, and the opposite direction will be considered downward, but these directions are unrelated to the direction of gravity. Viewing along the Z direction will be called a "planar view."

[0009] <Defect detection device and arc welding device> Figure 1 is a schematic perspective view showing a defect detection device and an arc welding device according to an embodiment. FIG. 2 is a block diagram schematically showing a defect detection device and an arc welding device according to an embodiment. FIG. 3 is a side view schematically showing the positions of cameras in the defect detection device and the arc welding device according to the embodiment. FIG. 4 is a plan view schematically showing the positions of cameras in the defect detection device and the arc welding device according to the embodiment. FIG. 5 is an explanatory diagram schematically showing a first image acquired by a first camera of the defect detection device and the arc welding device according to the embodiment. FIG. 6 is an explanatory diagram schematically showing a second image acquired by a second camera of the defect detection device and the arc welding device according to the embodiment.

[0010] As shown in FIGS. 1 and 2, an arc welding device 110 according to an embodiment is a device that joins a first welded material WM1 and a second welded material WM2 by arc welding. Hereinafter, the first welded material WM1 and the second welded material WM2 are collectively referred to as a welding target S.

[0011] The arc welding device 110 includes, for example, a device that joins the first welded material WM1 and the second welded material WM2 by TIG (Tungsten Inert Gas) welding, MIG (Metal Inert Gas) welding, MAG (Metal Active Gas) welding, carbon dioxide arc welding, or the like.

[0012] The first welded material WM1 has a flat plate shape. The first welded material WM1 is, for example, a heat sink. The first welded material WM1 is, for example, a panel joined to fins.

[0013] The second welded material WM2 includes, for example, a flat plate-shaped first portion and a second portion that is a standing portion rising from the first portion. The second portion extends, for example, in a direction orthogonal to the first portion. The second welded material WM2 is, for example, a fin. In FIG. 1, the second portion is omitted.

[0014] The arc welding apparatus 110 joins the second workpiece WM2 to the first workpiece WM1 by melting a welding rod 35 with a torch 30 and dripping it onto the first workpiece WM1 with the first portion of the second workpiece WM2 superimposed on it.

[0015] The arc welding apparatus 110 includes a defect detection device 100 and a torch 30. The defect detection device 100 will be described later. The arc welding apparatus 110 is capable of performing the arc welding method described later.

[0016] The torch 30 is provided with a metal electrode 31. The tip of the electrode 31 is exposed from the torch 30. When a voltage is applied between the electrode 31 and the workpiece S, an arc discharge occurs. Either the electrode 31 or the workpiece S may be set to a common potential (e.g., ground potential), and only the potential of the other electrode 31 or workpiece S may be controlled.

[0017] The welding rod 35 is, for example, a metal wire. The tip of the welding rod 35 is placed in the space where an arc discharge is occurring. The arc discharge melts the tip of the welding rod 35, and it drips onto the object to be welded S. A molten pool MP is formed by the molten metal from the welding rod 35 and a portion of the object to be welded S that has been melted by this metal. The object to be welded S is welded as the molten welding rod 35 solidifies. When the molten pool MP solidifies, it becomes a weld bead.

[0018] The defect detection device 100 includes a camera 10 and a control device 20. The defect detection device 100 is capable of performing the defect detection method described later.

[0019] Camera 10 acquires images of the area around the molten pool MP during arc welding. Camera 10 is electrically connected to the control device 20. Camera 10 outputs the acquired images to the control device 20. In the example in Figure 1, camera 10 is mounted on an arm.

[0020] In the examples shown in Figures 1 and 2, camera 10 includes a first camera 11 and a second camera 12. The first camera 11 is positioned in front of the molten pool MP (+Y direction) in the welding direction (Y direction). The second camera 12 is positioned behind the molten pool MP (-Y direction) in the welding direction (Y direction).

[0021] As shown in Figure 3, the depression angle α of camera 10 is, for example, between 15° and 60°. In the example in Figure 3, the depression angle α1 of the first camera 11 is 30°. In the example in Figure 3, the depression angle α2 of the second camera 12 is 45°. The depression angle α is the angle between the line L1 connecting camera 10 and the molten pool MP and the horizontal line HL in a side view. The depression angle α1 of the first camera 11 is the angle between the line L1a connecting camera 11 and the molten pool MP and the horizontal line HL1. The depression angle α2 of the second camera 12 is the angle between the line L1b connecting camera 12 and the molten pool MP and the horizontal line HL2.

[0022] As shown in Figure 4, the angle β between the line L2 connecting camera 10 and molten pool MP in a plan view and the line L3 along the welding direction (Y direction) is, for example, between 0° and 60°. In the example in Figure 4, the angle β1 between the line L2a connecting first camera 11 and molten pool MP in a plan view and the line L3 along the welding direction (Y direction) is 20°. In the example in Figure 4, the angle β2 between the line L2b connecting second camera 12 and molten pool MP in a plan view and the line L3 along the welding direction (Y direction) is 35°.

[0023] In the arc welding apparatus 110, arc welding is performed while capturing images of the area around the molten pool MP using a camera 10. The images may be still images or moving images. The camera 10 includes, for example, a CCD (Charge-Coupled Device) image sensor or a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor. If necessary, the area around the molten pool MP may be illuminated by a lighting device while the images are being captured.

[0024] The control device 20 is capable of performing various calculations in the defect detection device 100 and the defect detection method. The control device 20 includes, for example, a CPU (Central Processing Unit). The control device 20 may, for example, perform all calculations in a single area, or it may be divided into multiple areas, each for each executable calculation.

[0025] In the arc welding apparatus 110, arc welding is performed while moving the torch 30 and welding rod 35 along the Y direction, which is the direction along which the boundary between the first workpiece WM1 and the second workpiece WM2 extends. In addition, by moving the camera 10 along the Y direction along with the torch 30 and welding rod 35, arc welding can be performed while continuously photographing the molten pool MP, which changes position along the Y direction.

[0026] Defects such as holes can occur in the molten pool MP during arc welding. These defects occur, for example, when the heat input to the first part of the second workpiece WM2 is too high, causing the first part to melt and fall off. If the molten pool MP solidifies with defects present, defects such as holes will remain in the weld bead. One way to suppress the occurrence of holes is to detect defects in the molten pool MP during welding and repair them. However, over-detection can occur when detecting defects. That is, when detecting defects, things that are not actually defects may be detected as defects.

[0027] Therefore, in the defect detection device 100 (arc welding device 110), the control device 20 detects the estimated defect region DR, which is estimated to be a defect, in the image acquired by the camera 10, and determines whether or not the estimated defect region DR is a defect.

[0028] First, the control device 20 detects the estimated defect region DR and molten pool MP in the image. The control device 20 detects the estimated defect region DR and molten pool MP based on a model created by machine learning, for example. The model is created by machine learning using, for example, images containing defects, images without defects, images containing molten pool MP, and images without molten pool MP as training data. The control device 20 may also detect the estimated defect region DR and molten pool MP by image processing, for example. In image processing, the estimated defect region DR and molten pool MP are detected based on brightness, for example. In image processing, for example, the portion of the image where the brightness is above a first threshold is determined to be a molten pool MP. In image processing, for example, the portion of the image where the brightness is above a second threshold and below the first threshold is determined to be an estimated defect region DR.

[0029] Next, the control device 20 determines whether the estimated defect region DR is a defect based on the positional relationship between the estimated defect region DR and the molten pool MP. More specifically, the control device 20 determines that the estimated defect region DR is a defect when it is in a predetermined position relative to the molten pool MP, and determines that the estimated defect region DR is not a defect when it is not in a predetermined position relative to the molten pool MP.

[0030] The positional relationship between the estimated defect region (DR) and the molten pool (MP) will be explained in more detail below using Figures 5 and 6. Figure 5 schematically represents the first image acquired by the first camera 11. Figure 6 schematically represents the second image acquired by the second camera 12.

[0031] The control device 20 detects the estimated defect region DR and the molten pool MP in the first image of Figure 5 and the second image of Figure 6. In the first image of Figure 5, the control device 20 detects the first estimated defect region DR1 (first estimated defect region DR1a and first estimated defect region DR1b) and the molten pool MP. In the second image of Figure 6, the control device 20 detects the second estimated defect region DR2 (second estimated defect region DR2a and second estimated defect region DR2b) and the molten pool MP. The control device 20 then determines that the estimated defect region DR is a defect when it is in a predetermined position relative to the molten pool MP, and determines that the estimated defect region DR is not a defect when it is not in a predetermined position relative to the molten pool MP.

[0032] The predetermined position is, for example, a position in contact with the molten pool MP. In this case, the control device 20 determines, for example, that the estimated defect region DR is a defect when it is in contact with the molten pool MP, and determines that the estimated defect region DR is not a defect when it is not in contact with the molten pool MP.

[0033] In the example shown in Figure 5, since the first estimated defect region DR1a is in contact with the molten pool MP, the control device 20 determines that the first estimated defect region DR1a is a defect. Also in the example shown in Figure 5, since the first estimated defect region DR1b is not in contact with the molten pool MP, the control device 20 determines that the first estimated defect region DR1b is not a defect.

[0034] In the example shown in Figure 6, since the second estimated defect region DR2a is in contact with the molten pool MP, the control device 20 determines that the second estimated defect region DR2a is a defect. Also in the example shown in Figure 6, since the second estimated defect region DR2b is not in contact with the molten pool MP, the control device 20 determines that the second estimated defect region DR2b is not a defect.

[0035] The predetermined position may be, for example, a position where the angle γ between a straight line L4 connecting the center CD of the estimated defect region DR and the center CM of the molten pool MP, and a straight line L5 extending from the center CM of the molten pool MP in the direction of welding (Y direction) forward in the direction of welding (+Y direction), is less than a threshold. In this case, the control device 20 determines, for example, that the estimated defect region DR is defective when the angle γ is less than the threshold, and determines that the estimated defect region DR is not defective when the angle γ is greater than or equal to the threshold.

[0036] In the example in Figure 5, the threshold for angle γ is, for example, 0° or more and 50° or less. In the example in Figure 5, since the angle γ1a between the line L4a connecting the center CD1a of the first estimated defect region DR1a and the center CM of the molten pool MP, and the line L5, is less than 50°, the control device 20 determines that the first estimated defect region DR1a is defective. Also in the example in Figure 5, since the angle γ1b between the line L4b connecting the center CD1b of the first estimated defect region DR1b and the center CM of the molten pool MP, and the line L5, is 50° or more, the control device 20 determines that the first estimated defect region DR1b is not defective.

[0037] In the example in Figure 6, the threshold for angle γ is, for example, 0° or more and 80° or less. In the example in Figure 6, since the angle γ2a between the line L4a connecting the center CD2a of the second estimated defect region DR2a and the center CM of the molten pool MP, and the line L5, is less than 80°, the control device 20 determines that the second estimated defect region DR2a is defective. Also in the example in Figure 6, since the angle γ2b between the line L4b connecting the center CD2b of the second estimated defect region DR2b and the center CM of the molten pool MP, and the line L5, is 80° or more, the control device 20 determines that the second estimated defect region DR2b is not defective.

[0038] The threshold for angle γ is calculated, for example, from the threshold for the defect occurrence angle in a plan view and the shooting position of each image. The defect occurrence angle in a plan view is the angle between the line L4 and line L5 in a plan view. The threshold for the defect occurrence angle in a plan view is determined based on at least one of numerical analysis and statistical data analysis. The threshold for the defect occurrence angle in a plan view is, for example, between 0° and 60°. The threshold for angle γ in the first image is calculated from the threshold for the defect occurrence angle in a plan view and the shooting position of the first image (i.e., the position of the first camera 11). The threshold for angle γ in the first image is the value obtained by coordinate transformation of the threshold for the defect occurrence angle in a plan view (0° and 60°) from the plan view to the shooting position of the first image (i.e., the position of the first camera 11). The threshold for angle γ in the second image is calculated from the threshold for the defect occurrence angle in a plan view and the shooting position of the second image (i.e., the position of the second camera 12). The threshold angle γ in the second image is the value obtained by transforming the coordinates of the threshold angle for defect occurrence in the plan view (0° to 60°) from the plan view to the position where the second image was captured (i.e., the position of the second camera 12).

[0039] Furthermore, the control device 20 may determine whether the estimated defect area DR is a defect based, for example, on the positional relationship between the estimated defect area DR and the molten pool MP, and the elapsed time since the estimated defect area DR was first detected. Alternatively, the control device 20 may determine whether the estimated defect area DR is a defect based on an image taken from one direction (i.e., one of the first and second images), or on images taken from two directions (i.e., both the first and second images), or on images taken from three or more directions. The determination of whether the estimated defect area DR is a defect will be described later.

[0040] The control device 20 adjusts the arc welding conditions based on the judgment result. The control device 20 is electrically connected to the torch 30. The control device 20 adjusts the arc welding conditions, for example, the current value, voltage value, and energizing time applied to the torch 30. For example, the control device 20 can increase the amount of melted welding rod 35 and repair defects by increasing the current value (voltage value). Also, for example, the control device 20 can increase the amount of melted welding rod 35 and repair defects by increasing the energizing time.

[0041] Furthermore, the control device 20 may adjust the position of the welding rod 35 in arc welding as a condition for arc welding. The control device 20 can repair a defect by, for example, moving the welding rod 35 in the opposite direction of travel (-Y direction) in the Y direction, thereby melting the welding rod 35 near the defect. In other words, the control device 20 can repair a defect by moving the welding rod 35 closer to the defect. Alternatively, the control device 20 can repair a defect by, for example, moving the welding rod 35 in the X direction to approach the defect, thereby melting the welding rod 35 near the defect.

[0042] Furthermore, the control device 20 may also adjust the conditions for arc welding, for example, the movement speed of the welding rod 35 during arc welding. For example, by slowing down the movement speed of the welding rod 35, the control device 20 can melt the welding rod 35 near the defect and repair the defect.

[0043] As described above, the defect detection device 100 and arc welding apparatus 110 according to this embodiment detect the estimated defect area DR and the molten pool MP, and determine whether the estimated defect area DR is a defect or not based on the positional relationship between the estimated defect area DR and the molten pool MP. For example, it is possible to determine that any estimated defect area DR that is not in a predetermined position relative to the molten pool MP is not a defect. This suppresses false positives and enables accurate defect detection.

[0044] Furthermore, in the arc welding apparatus 110 according to this embodiment, defects that occur during arc welding can be repaired during arc welding by adjusting the arc welding conditions based on the judgment result. This makes it possible to suppress the occurrence of perforation defects.

[0045] Furthermore, in the defect detection device 100 and arc welding apparatus 110 according to the embodiment, the camera 10 includes a first camera 11 positioned in front of the molten pool MP and a second camera 12 positioned behind the molten pool MP. This allows the second camera 12 to capture, for example, the estimated defect area DR, which is difficult to capture from the first camera 11 because it is hidden in the shadow of the second part of the second workpiece material WM2. This reduces the chance of overlooking the estimated defect area DR and enables more accurate defect detection.

[0046] Furthermore, in the defect detection device 100 and arc welding device 110 according to the embodiment, by setting the depression angle of the camera 10 to 15° or more and 60° or less, it is possible to further suppress overlooking the estimated defect area DR and detect defects with greater accuracy.

[0047] Furthermore, in the defect detection device 100 and arc welding device 110 according to the embodiment, by setting the angle between the straight line L4 connecting the camera 10 and the molten pool MP and the straight line L5 along the welding direction to 0° or more and 60° or less in a plan view, it is possible to further suppress overlooking the estimated defect area DR and detect defects with greater accuracy.

[0048] <Defect detection method> Figure 7 is a flowchart showing an example of a defect detection method according to the first embodiment. As shown in Figure 7, the defect detection method according to the first embodiment includes steps S101 to S105. Step S101 corresponds to the acquisition step. Step S102 corresponds to the detection step. Steps S103 to S105 correspond to the determination step.

[0049] When arc welding is started, the control device 20 first acquires an image of the area around the molten pool MP during arc welding (step S101). The image is acquired by the camera 10.

[0050] The control device 20 then detects the estimated defect region DR and molten pool MP in the image (step S102). The control device 20 detects the estimated defect region DR and molten pool MP based, for example, on a model created by machine learning.

[0051] The control device 20 then determines whether the estimated defect region DR is a defect based on the positional relationship between the estimated defect region DR and the molten pool MP. More specifically, the control device 20 determines whether the estimated defect region DR is at a predetermined position relative to the molten pool MP (step S103). The predetermined position is as described above.

[0052] When the estimated defect region DR is in a predetermined position relative to the molten pool MP (step S103: Yes), the control device 20 determines that the estimated defect region DR is a defect (step S104).

[0053] If the estimated defect region DR is not in a predetermined position relative to the molten pool MP (step S103: No), the control device 20 determines that the estimated defect region DR is not a defect (step S105).

[0054] In other words, in the defect detection method according to the first embodiment, the control device 20 determines that the estimated defect region DR is a defect when it is located at a predetermined position relative to the molten pool MP. The control device 20 determines that the estimated defect region DR is not a defect when it is not located at a predetermined position relative to the molten pool MP.

[0055] As described above, the defect detection method according to the first embodiment detects the estimated defect region DR and the molten pool MP, and determines whether the estimated defect region DR is a defect based on the positional relationship between the estimated defect region DR and the molten pool MP. For example, it is possible to determine that any estimated defect region DR that is not in a predetermined position relative to the molten pool MP is not a defect. This suppresses false positives and enables accurate defect detection.

[0056] In this example, the control device 20 detects one estimated defect region DR from one image during the detection process. However, the control device 20 may detect multiple estimated defect regions DR from a single image during the detection process. In this case, the control device 20 determines whether or not each of the multiple estimated defect regions DR is a defect during the determination process.

[0057] Figure 8 is a flowchart showing an example of a defect detection method according to the second embodiment. As shown in Figure 8, the defect detection method according to the second embodiment is substantially the same as the defect detection method according to the first embodiment, except that in the determination step, it determines whether or not the estimated defect region DR is a defect based on the positional relationship between the estimated defect region DR and the molten pool MP, and the elapsed time since the estimated defect region DR began to be detected.

[0058] The defect detection method according to the second embodiment includes steps S201 to S206. Step S201 corresponds to the acquisition step. Step S202 corresponds to the detection step. Steps S203 to S206 correspond to the determination step. Steps S201 to S203 of the defect detection method according to the second embodiment are substantially the same as steps S101 to S103 of the defect detection method according to the first embodiment.

[0059] In the determination step, the control device 20 determines whether the estimated defect area DR is in a predetermined position relative to the molten pool MP (step S203). If the estimated defect area DR is in a predetermined position relative to the molten pool MP (step S203: Yes), the control device 20 determines whether a predetermined time has elapsed since the estimated defect area DR was first detected (step S204).

[0060] When a predetermined time has elapsed since the estimated defect area DR began to be detected (step S204: Yes), the control device 20 determines that the estimated defect area DR is a defect (step S205). The predetermined time is, for example, 0.5 seconds or more (for example, 0.6 seconds).

[0061] If a predetermined amount of time has not elapsed since the estimated defect area DR began to be detected (step S204: No), the control device 20 determines that the estimated defect area DR is not a defect (step S206).

[0062] If the estimated defect region DR is not in a predetermined position relative to the molten pool MP (step S203: No), the control device 20 determines that the estimated defect region DR is not a defect (step S206).

[0063] In other words, in the defect detection method according to the second embodiment, the control device 20 determines that the estimated defect region DR is a defect when the estimated defect region DR is in a predetermined position relative to the molten pool MP and a predetermined time has elapsed since the estimated defect region DR began to be detected. The control device 20 determines that the estimated defect region DR is not a defect when the estimated defect region DR is in a predetermined position relative to the molten pool MP and a predetermined time has not elapsed since the estimated defect region DR began to be detected. The control device 20 determines that the estimated defect region DR is not a defect when the estimated defect region DR is not in a predetermined position relative to the molten pool MP.

[0064] As described above, the defect detection method according to the second embodiment detects the estimated defect region DR and the molten pool MP, and determines whether the estimated defect region DR is a defect based on the positional relationship between the estimated defect region DR and the molten pool MP, and the elapsed time since the estimated defect region DR began to be detected. For example, it is possible to determine that estimated defect regions DR that are not in a predetermined position relative to the molten pool MP or that have not been detected for a predetermined time are not defects. This suppresses false positives and enables more accurate defect detection.

[0065] Step S204 may be performed before step S203, or it may be performed simultaneously with step S203.

[0066] In this example, the control device 20 detects one estimated defect region DR from one image during the detection process. However, the control device 20 may detect multiple estimated defect regions DR from a single image during the detection process. In this case, the control device 20 determines whether or not each of the multiple estimated defect regions DR is a defect during the determination process.

[0067] Figure 9 is a flowchart showing an example of a defect detection method according to the third embodiment. As shown in Figure 9, the defect detection method according to the third embodiment is substantially the same as the defect detection method according to the first embodiment, except that in the acquisition step, a first image and a second image are acquired; in the detection step, a first estimated defect region DR1 and a molten pool MP are detected in the first image, and a second estimated defect region DR2 and a molten pool MP are detected in the second image; and in the determination step, it is determined whether the first estimated defect region DR1 is a defect or not based on the positional relationship between the first estimated defect region DR1 and the molten pool MP, and whether the second estimated defect region DR2 is a defect or not based on the positional relationship between the second estimated defect region DR2 and the molten pool MP. In the defect detection method according to the third embodiment, the control device 20 determines whether the first estimated defect region DR1 and the second estimated defect region DR2 are defects or not without determining whether the first estimated defect region DR1 is the same as the second estimated defect region DR2.

[0068] The defect detection method according to the third embodiment includes steps S301 to S308. Step S301 corresponds to the acquisition step. Step S302 corresponds to the detection step. Steps S303 to S308 correspond to the determination step. Steps S301 to S303 of the defect detection method according to the third embodiment are substantially the same as steps S101 to S103 of the defect detection method according to the first embodiment.

[0069] In the acquisition process, the control device 20 acquires a first image and a second image of the area around the molten pool MP during arc welding (step S301).

[0070] In the detection process, the control device 20 detects the first estimated defect region DR1 and the molten pool MP in the first image, and detects the second estimated defect region DR2 and the molten pool MP in the second image (step S302).

[0071] In the determination process, the control device 20 determines whether the first estimated defect region DR1 is in a predetermined position relative to the molten pool MP (step S303).

[0072] When the first estimated defect region DR1 is in a predetermined position relative to the molten pool MP (step S303: Yes), the control device 20 determines that the first estimated defect region DR1 is a defect (step S304).

[0073] If the first estimated defect area DR1 is not in a predetermined position relative to the molten pool MP (step S303: No), the control device 20 determines that the first estimated defect area DR1 is not a defect (step S305).

[0074] The control device 20 then determines whether the second estimated defect region DR2 is in a predetermined position relative to the molten pool MP (step S306).

[0075] When the second estimated defect region DR2 is in a predetermined position relative to the molten pool MP (step S306: Yes), the control device 20 determines that the second estimated defect region DR2 is a defect (step S307).

[0076] If the second estimated defect region DR2 is not in a predetermined position relative to the molten pool MP (step S306: No), the control device 20 determines that the second estimated defect region DR2 is not a defect (step S308).

[0077] In other words, in the defect detection method according to the third embodiment, the control device 20 determines that the first estimated defect region DR1 is a defect when it is located at a predetermined position relative to the molten pool MP. The control device 20 determines that the first estimated defect region DR1 is not a defect when it is not located at a predetermined position relative to the molten pool MP. The control device 20 determines that the second estimated defect region DR2 is a defect when it is located at a predetermined position relative to the molten pool MP. The control device 20 determines that the second estimated defect region DR2 is not a defect when it is not located at a predetermined position relative to the molten pool MP.

[0078] As described above, the defect detection method according to the third embodiment detects a first estimated defect region DR1, a second estimated defect region DR2, and a molten pool MP. Based on the positional relationship between the first estimated defect region DR1 and the molten pool MP, it is determined whether or not the first estimated defect region DR1 is a defect. Based on the positional relationship between the second estimated defect region DR2 and the molten pool MP, it is determined whether or not the second estimated defect region DR2 is a defect. For example, among the first and second estimated defect regions DR1 and DR2, those that are not in a predetermined position relative to the molten pool MP can be determined not to be defects. This suppresses false positives and enables accurate defect detection.

[0079] Steps S306 to S308 may be performed before steps S303 to S305, or they may be performed simultaneously with steps S303 to S305.

[0080] In this example, the control device 20 detects one estimated defect region DR from one image in the detection step, but the control device 20 may detect multiple estimated defect regions DR from one image in the detection step. In this case, the control device 20 makes a determination in the determination step to determine whether or not each of the multiple estimated defect regions DR is a defect. Also, in this example, the control device 20 acquires two images in the acquisition step, but the control device 20 may acquire three or more images in the acquisition step. In this case, the control device 20 detects estimated defect regions DR from each of the three or more images in the detection step, and makes a determination in the determination step to determine whether or not each of the multiple estimated defect regions DR is a defect.

[0081] Figure 10 is a flowchart showing an example of a defect detection method according to the fourth embodiment. As shown in Figure 10, the defect detection method according to the fourth embodiment is substantially the same as the defect detection method according to the third embodiment, except that in the determination step, it determines whether the first estimated defect area DR1 is a defect based on the positional relationship between the first estimated defect area DR1 and the molten pool MP, and the elapsed time since the first estimated defect area DR1 was detected, and determines whether the second estimated defect area DR2 is a defect based on the positional relationship between the second estimated defect area DR2 and the molten pool MP, and the elapsed time since the second estimated defect area DR2 was detected. In the defect detection method according to the fourth embodiment, the control device 20 determines whether each of the first estimated defect area DR1 and the second estimated defect area DR2 is a defect without determining whether the first estimated defect area DR1 is the same as the second estimated defect area DR2.

[0082] The defect detection method according to the fourth embodiment includes steps S401 to S410. Step S401 corresponds to the acquisition step. Step S402 corresponds to the detection step. Steps S403 to S410 correspond to the determination step. Steps S401 to S403 and S407 of the defect detection method according to the fourth embodiment are substantially the same as steps S301 to S303 and S306 of the defect detection method according to the third embodiment.

[0083] In the determination step, the control device 20 determines whether the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP (step S403). If the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP (step S403: Yes), the control device 20 determines whether a predetermined time has elapsed since the first estimated defect area DR1 was first detected (step S404).

[0084] When a predetermined time has elapsed since the first estimated defect region DR1 began to be detected (step S404: Yes), the control device 20 determines that the first estimated defect region DR1 is a defect (step S405). The predetermined time is, for example, 0.5 seconds or more (for example, 0.6 seconds).

[0085] If a predetermined amount of time has not elapsed since the detection of the first estimated defect region DR1 began (step S404: No), the control device 20 determines that the first estimated defect region DR1 is not a defect (step S406).

[0086] If the first estimated defect region DR1 is not in a predetermined position relative to the molten pool MP (step S403: No), the control device 20 determines that the first estimated defect region DR1 is not a defect (step S406).

[0087] The control device 20 then determines whether the second estimated defect region DR2 is in a predetermined position relative to the molten pool MP (step S407). If the second estimated defect region DR2 is in a predetermined position relative to the molten pool MP (step S407: Yes), the control device 20 determines whether a predetermined time has elapsed since the second estimated defect region DR2 was first detected (step S408).

[0088] When a predetermined time has elapsed since the detection of the second estimated defect region DR2 began (step S408: Yes), the control device 20 determines that the second estimated defect region DR2 is a defect (step S409). The predetermined time is, for example, 0.5 seconds or more (for example, 0.6 seconds).

[0089] If a predetermined amount of time has not elapsed since the detection of the second estimated defect region DR2 began (step S408: No), the control device 20 determines that the second estimated defect region DR2 is not a defect (step S410).

[0090] If the second estimated defect region DR2 is not in a predetermined position relative to the molten pool MP (step S407: No), the control device 20 determines that the second estimated defect region DR2 is not a defect (step S410).

[0091] In other words, in the defect detection method according to the fourth embodiment, the control device 20 determines that the first estimated defect area DR1 is a defect when the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP and a predetermined time has elapsed since the first estimated defect area DR1 began to be detected. The control device 20 determines that the first estimated defect area DR1 is not a defect when the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP and a predetermined time has not elapsed since the first estimated defect area DR1 began to be detected. The control device 20 determines that the first estimated defect area DR1 is not a defect when the first estimated defect area DR1 is not in a predetermined position relative to the molten pool MP. The control device 20 determines that the second estimated defect area DR2 is a defect when the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP and a predetermined time has elapsed since the second estimated defect area DR2 began to be detected. The control device 20 determines that the second estimated defect area DR2 is not a defect if the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP and a predetermined time has not elapsed since the second estimated defect area DR2 was first detected. The control device 20 determines that the second estimated defect area DR2 is not a defect if the second estimated defect area DR2 is not in a predetermined position relative to the molten pool.

[0092] As described above, the defect detection method according to the fourth embodiment detects a first estimated defect area DR1, a second estimated defect area DR2, and a molten pool MP. Based on the positional relationship between the first estimated defect area DR1 and the molten pool MP, and the elapsed time since the first estimated defect area DR1 was detected, it is determined whether or not the first estimated defect area DR1 is a defect. Similarly, based on the positional relationship between the second estimated defect area DR2 and the molten pool MP, and the elapsed time since the second estimated defect area DR2 was detected, it is determined whether or not the second estimated defect area DR2 is a defect. For example, among the first and second estimated defect areas DR1 and DR2, those that are not in a predetermined position relative to the molten pool MP or those that have not been detected for a predetermined amount of time can be determined not to be defects. This suppresses false positives and enables more accurate defect detection.

[0093] Step S404 may be performed before step S403, or simultaneously with step S403. Also, step S408 may be performed before step S407, or simultaneously with step S407. Furthermore, steps S407 to S410 may be performed before steps S403 to S406, or simultaneously with steps S403 to S406.

[0094] In this example, the control device 20 detects one estimated defect region DR from one image in the detection step, but the control device 20 may detect multiple estimated defect regions DR from one image in the detection step. In this case, the control device 20 makes a determination in the determination step to determine whether or not each of the multiple estimated defect regions DR is a defect. Also, in this example, the control device 20 acquires two images in the acquisition step, but the control device 20 may acquire three or more images in the acquisition step. In this case, the control device 20 detects estimated defect regions DR from each of the three or more images in the detection step, and makes a determination in the determination step to determine whether or not each of the multiple estimated defect regions DR is a defect.

[0095] Figure 11 is a flowchart showing an example of a defect detection method according to the fifth embodiment. As shown in Figure 11, the defect detection method according to the fifth embodiment is substantially the same as the defect detection method according to the third embodiment, except that in the detection step, it estimates whether the first estimated defect area DR1 is the same as the second estimated defect area DR2, performs a first determination process if the first estimated defect area DR1 is the same as the second estimated defect area DR2, and performs a second determination process if the first estimated defect area DR1 is not the same as the second estimated defect area DR2.

[0096] The defect detection method according to the fifth embodiment includes steps S501 to S505. Step S501 corresponds to the acquisition step. Steps S502 and S503 correspond to the detection step. Steps S504 and S505 correspond to the determination step. Steps S501 and S502 of the defect detection method according to the fifth embodiment are substantially the same as steps S301 and S302 of the defect detection method according to the third embodiment.

[0097] In the detection process, the control device 20 estimates whether the first estimated defect area DR1 is the same as the second estimated defect area DR2 (step S503). For example, the control device 20 estimates that the first estimated defect area DR1 is the same as the second estimated defect area DR2 if the positional relationship between the first estimated defect area DR1 and the molten pool MP is the same as the positional relationship between the second estimated defect area DR2 and the molten pool MP. For example, the control device 20 estimates that the first estimated defect area DR1 is not the same as the second estimated defect area DR2 if the positional relationship between the first estimated defect area DR1 and the molten pool MP is different from the positional relationship between the second estimated defect area DR2 and the molten pool MP.

[0098] More specifically, the control device 20 estimates the absolute coordinate positions of the first estimated defect region DR1 and the molten pool MP based, for example, the angle γ in the first image shown in Figure 5 and the position where the first image was taken (i.e., the position of the first camera 11). The control device 20 estimates the absolute coordinate positions of the second estimated defect region DR2 and the molten pool MP based, for example, the angle γ in the second image shown in Figure 6 and the position where the second image was taken (i.e., the position of the second camera 12). The control device 20 estimates that the first estimated defect region DR1 is the same as the second estimated defect region DR2 when the absolute coordinate positions of the first estimated defect region DR1 and the molten pool MP are the same as the absolute coordinate positions of the second estimated defect region DR2 and the molten pool MP.

[0099] In the determination process, the control device 20 performs either a first determination process or a second determination process based on the estimation result. More specifically, when it is estimated that the first estimated defect area DR1 is the same as the second estimated defect area DR2 (step S503: Yes), the control device 20 performs the first determination process (step S504). When it is estimated that the first estimated defect area DR1 is not the same as the second estimated defect area DR2 (step S503: No), the control device 20 performs the second determination process (step S505).

[0100] Figure 12 is a flowchart showing an example of the first determination process of the defect detection method according to the fifth embodiment. As shown in Figure 12, in the first determination process, the control device 20 determines whether the estimated defect region DRX, which is both the first estimated defect region DR1 and the second estimated defect region DR2, is a defect, based on the premise that the first estimated defect region DR1 is the same as the second estimated defect region DR2. The first determination process includes steps S601 to S604.

[0101] In the first determination process, the control device 20 determines whether at least one of the first estimated defect region DR1 and the second estimated defect region DR2 is located in a predetermined position relative to the molten pool MP (step S601).

[0102] When at least one of the first estimated defect area DR1 and the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP (step S601: Yes), the control device 20 determines whether a predetermined time has elapsed since at least one of the first estimated defect area DR1 and the second estimated defect area DR2 began to be detected (step S602). For example, the control device 20 starts counting time when at least one of the first estimated defect area DR1 and the second estimated defect area DR2 begins to be detected, and continues counting time until both the first estimated defect area DR1 and the second estimated defect area DR2 are no longer detected. In other words, for example, the control device 20 continues counting time not only while one of the first estimated defect area DR1 and the second estimated defect area DR2 is detected, but also while one of the first estimated defect area DR1 and the second estimated defect area DR2 is detected and the other is not detected.

[0103] When a predetermined time has elapsed since at least one of the first estimated defect region DR1 and the second estimated defect region DR2 began to be detected (step S602: Yes), the control device 20 determines that the estimated defect region DRX is a defect (step S603).

[0104] If a predetermined amount of time has not elapsed since at least one of the first estimated defect region DR1 and the second estimated defect region DR2 began to be detected (step S602: No), the control device 20 determines that the estimated defect region DRX is not a defect (step S604).

[0105] If both the first estimated defect region DR1 and the second estimated defect region DR2 are not in a predetermined position relative to the molten pool MP (step S601: No), the control device 20 determines that the estimated defect region DRX is not a defect (step S604).

[0106] In other words, in the first determination process, the control device 20 determines that the first estimated defect area DR1 and the second estimated defect area DR2 (estimated defect area DRX) are defects if at least one of the first estimated defect area DR1 and the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP, and a predetermined time has elapsed since at least one of the first estimated defect area DR1 and the second estimated defect area DR2 began to be detected. The control device 20 determines that the first estimated defect area DR1 and the second estimated defect area DR2 (estimated defect area DRX) are not defects if at least one of the first estimated defect area DR1 and the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP, and a predetermined time has not elapsed since at least one of the first estimated defect area DR1 and the second estimated defect area DR2 began to be detected. The control device 20 determines that the first estimated defect area DR1 and the second estimated defect area DR2 (estimated defect area DRX) are not defects if neither the first estimated defect area DR1 nor the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP.

[0107] Step S602 may be performed before step S601, or it may be performed simultaneously with step S601.

[0108] Figure 13 is a flowchart showing an example of the second determination process of the defect detection method according to the fifth embodiment. As shown in Figure 12, in the second determination process, the control device 20 determines whether the first estimated defect region DR1 and the second estimated defect region DR2 are defects, based on the premise that the first estimated defect region DR1 is not the same as the second estimated defect region DR2.

[0109] The second determination process is substantially the same as the determination step of the defect detection method according to the fourth embodiment. The second determination process includes steps S701 to S708. Steps S701 to S708 of the second determination process are substantially the same as steps S403 to S410 of the defect detection method according to the fourth embodiment.

[0110] In other words, in the second determination process, the control device 20 determines that the first estimated defect area DR1 is a defect when the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP and a predetermined time has elapsed since the first estimated defect area DR1 was first detected. The control device 20 determines that the first estimated defect area DR1 is not a defect when the first estimated defect area DR1 is in a predetermined position relative to the molten pool MP and a predetermined time has not elapsed since the first estimated defect area DR1 was first detected. The control device 20 determines that the first estimated defect area DR1 is not a defect when the first estimated defect area DR1 is not in a predetermined position relative to the molten pool MP. The control device 20 determines that the second estimated defect area DR2 is a defect when the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP and a predetermined time has elapsed since the second estimated defect area DR2 was first detected. The control device 20 determines that the second estimated defect area DR2 is not a defect if the second estimated defect area DR2 is in a predetermined position relative to the molten pool MP and a predetermined time has not elapsed since the second estimated defect area DR2 was first detected. The control device 20 determines that the second estimated defect area DR2 is not a defect if the second estimated defect area DR2 is not in a predetermined position relative to the molten pool MP.

[0111] Thus, in the defect detection method according to the fifth embodiment, a first estimated defect region DR1, a second estimated defect region DR2, and a molten pool MP are detected, and it is estimated whether the first estimated defect region DR1 is the same as the second estimated defect region DR2. If the first estimated defect region DR1 is the same as the second estimated defect region DR2, a first determination process is performed, allowing the time count to continue even if either the first estimated defect region DR1 or the second estimated defect region DR2 is no longer visible in the image. This allows for accurate detection of the elapsed time since the first estimated defect region DRX, which is both the first estimated defect region DR1 and the second estimated defect region DR2, was first detected, and thus enables accurate defect detection. Furthermore, if the first estimated defect region DR1 is not the same as the second estimated defect region DR2, a second determination process is performed, enabling accurate defect detection.

[0112] Furthermore, when the positional relationship between the first estimated defect area DR1 and the molten pool MP is the same as the positional relationship between the second estimated defect area DR2 and the molten pool MP, it is estimated that the first estimated defect area DR1 is identical to the second estimated defect area DR2. Conversely, when the positional relationship between the first estimated defect area DR1 and the molten pool MP is different from the positional relationship between the second estimated defect area DR2 and the molten pool MP, it is estimated that the first estimated defect area DR1 is not identical to the second estimated defect area DR2. This allows for an accurate estimation of whether or not the first estimated defect area DR1 is identical to the second estimated defect area DR2.

[0113] Step S702 may be performed before step S701, or simultaneously with step S701. Also, step S706 may be performed before step S705, or simultaneously with step S705. Furthermore, steps S705 to S708 may be performed before steps S701 to S704, or simultaneously with steps S701 to S704.

[0114] In this example, the control device 20 detects one estimated defect region DR from one image in the detection step, but the control device 20 may detect multiple estimated defect regions DR from one image in the detection step. In this case, the control device 20 makes a determination in the determination step to determine whether or not each of the multiple estimated defect regions DR is a defect. Also, in this example, the control device 20 acquires two images in the acquisition step, but the control device 20 may acquire three or more images in the acquisition step. In this case, the control device 20 detects estimated defect regions DR from each of the three or more images in the detection step, estimates whether or not the multiple estimated defect regions DR are the same, and in the determination step performs at least one of the first determination process and the second determination process based on the estimation result.

[0115] <Arc welding method> Figure 14 is a flowchart illustrating an example of an arc welding method according to the embodiment. As shown in Figure 14, the arc welding method according to the embodiment includes steps S801 to S807. Step S801 corresponds to the acquisition step. Step S802 corresponds to the detection step. Steps S803 to S805 correspond to the determination step. Steps S806 and S807 correspond to the adjustment step. Steps S801 to S805 of the arc welding method according to the embodiment are substantially the same as steps S101 to S105 of the defect detection method according to the first embodiment. Steps S801 to S805 of the arc welding method according to the embodiment may also be replaced with steps S201 to S206 of the defect detection method according to the second embodiment, steps S301 to S308 of the defect detection method according to the third embodiment, steps S401 to S410 of the defect detection method according to the fourth embodiment, steps S501 to S505 of the defect detection method according to the fifth embodiment, etc.

[0116] In the adjustment process, the control device 20 adjusts the arc welding conditions based on the judgment result. More specifically, the control device 20 changes the arc welding conditions (step S806) when the estimated defect area DR is a defect (step S804). The control device 20 maintains the arc welding conditions (step S807) when the estimated defect area DR is not a defect (step S805). The arc welding conditions are as described above.

[0117] As described above, in the arc welding method according to this embodiment, the estimated defect area DR and molten pool MP are detected, and based on the positional relationship between the estimated defect area DR and the molten pool MP, it is determined whether or not the estimated defect area DR is a defect. Based on the determination result, the arc welding conditions are adjusted, thereby repairing defects that occur during arc welding. This makes it possible to suppress the occurrence of perforated defects.

[0118] <Program and recording medium> The above-described defect detection method may be recorded as a program that can be executed by a computer on a magnetic disk (such as a flexible disk and a hard disk), an optical disk (such as a CD-ROM, CD-R, CD-RW, DVD-ROM, DVD±R, DVD±RW), a semiconductor memory, or another non-transitory computer-readable storage medium.

[0119] For example, information recorded on a recording medium can be read by a computer (or embedded system). The recording format (storage format) of the recording medium is arbitrary. For example, a computer reads a program from the recording medium and has the CPU execute the instructions written in the program based on this program. In a computer, program acquisition (or reading) may be performed via a network.

[0120] The embodiment may include the following configurations.

[0121] (Composition 1) The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Equipped with, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. A defect detection method that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

[0122] (Configuration 2) The defect detection method according to configuration 1, wherein the predetermined position is a position in contact with the molten pool.

[0123] (Composition 3) The defect detection method according to configuration 1 or 2, wherein the predetermined position is a position where the angle between a straight line connecting the center of the estimated defect region and the center of the molten pool and a straight line extending forward in the direction of welding from the center of the molten pool along the direction of welding is less than a threshold.

[0124] (Composition 4) In the determination step, When the estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the estimated defect region was first detected, it is determined that the estimated defect region is a defect. If the estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the estimated defect region was first detected, it is determined that the estimated defect region is not a defect. A defect detection method according to any one of configurations 1 to 3, wherein if the estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the estimated defect region is not a defect.

[0125] (Composition 5) In the acquisition step described above, a first image and a second image taken simultaneously with the first image from a different position than the first image are acquired. In the detection step, the first estimated defect region and the molten pool are detected in the first image, and the second estimated defect region and the molten pool are detected in the second image. In the determination step, When the first estimated defect region is located at the predetermined position relative to the molten pool, it is determined that the first estimated defect region is a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, it is determined that the second estimated defect region is a defect. A defect detection method according to any one of configurations 1 to 3, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

[0126] (Composition 6) In the determination step, When the first estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the first estimated defect region was first detected, it is determined that the first estimated defect region is a defect. If the first estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the first estimated defect region was detected, it is determined that the first estimated defect region is not a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is a defect. If the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is not a defect. The defect detection method according to configuration 5, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

[0127] (Composition 7) In the determination step, Based on a threshold calculated from the position where the first image was captured, it is determined whether the first estimated defect region is located at the predetermined position relative to the molten pool. A defect detection method according to configuration 5 or 6, which determines whether the second estimated defect region is in the predetermined position relative to the molten pool based on a threshold calculated from the position of the second image.

[0128] (Composition 8) In the acquisition step described above, a first image and a second image taken simultaneously with the first image from a different position than the first image are acquired. In the detection step, In the first image, the first estimated defect region and the molten pool are detected, and in the second image, the second estimated defect region and the molten pool are detected. We estimate whether the first estimated defect region is the same as the second estimated defect region. When the first estimated defect region is estimated to be the same as the second estimated defect region, the first determination process is performed in the determination step. When it is determined that the first estimated defect region is not the same as the second estimated defect region, the determination process is performed, In the first determination process described above, When at least one of the first estimated defect region and the second estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since at least one of the first estimated defect region and the second estimated defect region began to be detected, it is determined that the first estimated defect region and the second estimated defect region are defects. If at least one of the first estimated defect region and the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since at least one of the first estimated defect region and the second estimated defect region began to be detected, then it is determined that the first estimated defect region and the second estimated defect region are not defects. When both the first estimated defect region and the second estimated defect region are not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region and the second estimated defect region are not defects. In the second determination process described above, When the first estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the first estimated defect region was first detected, it is determined that the first estimated defect region is a defect. If the first estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the first estimated defect region was detected, it is determined that the first estimated defect region is not a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is a defect. If the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is not a defect. A defect detection method according to any one of configurations 1 to 3, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

[0129] (Composition 9) In the detection step, When the positional relationship between the first estimated defect region and the molten pool is the same as the positional relationship between the second estimated defect region and the molten pool, the first estimated defect region is estimated to be identical to the second estimated defect region. The defect detection method according to configuration 8, wherein when the positional relationship between the first estimated defect region and the molten pool is different from the positional relationship between the second estimated defect region and the molten pool, it is estimated that the first estimated defect region is not the same as the second estimated defect region.

[0130] (Composition 10) The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Based on the judgment result, an adjustment step is made to adjust the conditions of the arc welding, Equipped with, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. An arc welding method in which, when the estimated defect region is not located in the predetermined position relative to the molten pool, it is determined that the estimated defect region is not a defect.

[0131] (Composition 11) A camera that acquires images of the area around the molten pool during arc welding, A control device that determines whether or not the estimated defect region in the aforementioned image is actually a defect, Equipped with, The control device is The estimated defect region and the molten pool are detected, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. A defect detection device that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

[0132] (Composition 12) The camera includes a first camera and a second camera, The first camera is positioned in front of the molten pool in the direction of welding, The defect detection device according to configuration 11, wherein the second camera is positioned behind the molten pool in the direction of welding.

[0133] (Composition 13) The defect detection device according to configuration 11, wherein the depression angle of the camera is 15° or more and 60° or less.

[0134] (Composition 14) The defect detection device according to configuration 11, wherein, in a plan view, the angle between the straight line connecting the camera and the molten pool and the straight line along the direction of welding is 0° or more and 60° or less.

[0135] (Composition 15) A camera that acquires images of the area around the molten pool during arc welding, A control device that determines whether the estimated defect area in the aforementioned image is actually a defect, and adjusts the arc welding conditions based on the determination result, Torch and, Equipped with, The control device is The estimated defect region and the molten pool are detected, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. An arc welding apparatus that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

[0136] (Composition 16) On the computer, The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Make it run, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, the estimated defect region is determined to be a defect. A program that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

[0137] (Composition 17) A recording medium on which the program described in configuration 16 is recorded.

[0138] As described above, according to the embodiment, it is possible to provide a defect detection method, an arc welding method, a defect detection device, an arc welding device, a program, and a recording medium that can detect defects with high accuracy.

[0139] The embodiments of the present invention have been illustrated above, but these embodiments are presented as examples only and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims of the invention and its equivalents. [Explanation of symbols]

[0140] 10: Camera 11: Camera 1 12: Second camera 20: Control device 30: Torch 31: Electrode 35: Welding rod 100: Defect detection device 110: Arc welding equipment CD, CD1a, CD1b, CD2a, CD2b, CM: Center DR, DRX: Estimated Defect Region DR1, DR1a, DR1b: First estimated defect region DR2, DR2a, DR2b: Second estimated defect region HL, HL1, HL2: Horizontal line L1, L1a, L1b, L2, L2a, L2b, L3, L4, L4a, L4b, L5: Straight line MP: Melt pool S: Welding target WM1: First material to be welded WM2: 2nd material to be welded α, α1, α2: Depression angle β, β1, β2, γ, γ1a, γ1b, γ2a, γ2b: Angle

Claims

1. The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Equipped with, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. A defect detection method that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

2. The defect detection method according to claim 1, wherein the predetermined position is a position in contact with the molten pool.

3. The defect detection method according to claim 1, wherein the predetermined position is a position where the angle between a straight line connecting the center of the estimated defect region and the center of the molten pool and a straight line extending from the center of the molten pool toward the front in the direction of welding, along the direction of welding, is less than a threshold.

4. In the determination step, When the estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the estimated defect region was first detected, it is determined that the estimated defect region is a defect. If the estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the estimated defect region was first detected, it is determined that the estimated defect region is not a defect. The defect detection method according to claim 1, wherein when the estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the estimated defect region is not a defect.

5. In the acquisition step described above, a first image and a second image taken simultaneously with the first image from a different position than the first image are acquired. In the detection step, the first estimated defect region and the molten pool are detected in the first image, and the second estimated defect region and the molten pool are detected in the second image. In the determination step, When the first estimated defect region is located at the predetermined position relative to the molten pool, it is determined that the first estimated defect region is a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, it is determined that the second estimated defect region is a defect. The defect detection method according to claim 1, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

6. In the determination step, When the first estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the first estimated defect region was first detected, it is determined that the first estimated defect region is a defect. If the first estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the first estimated defect region was detected, it is determined that the first estimated defect region is not a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is a defect. If the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is not a defect. The defect detection method according to claim 5, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

7. In the determination step, Based on a threshold calculated from the position where the first image was captured, it is determined whether the first estimated defect region is located at the predetermined position relative to the molten pool. The defect detection method according to claim 5, wherein it is determined whether the second estimated defect region is in the predetermined position relative to the molten pool based on a threshold calculated from the position of the second image.

8. In the acquisition step described above, a first image and a second image taken simultaneously with the first image from a different position than the first image are acquired. In the detection step, In the first image, the first estimated defect region and the molten pool are detected, and in the second image, the second estimated defect region and the molten pool are detected. We estimate whether the first estimated defect region is the same as the second estimated defect region. When the first estimated defect region is estimated to be the same as the second estimated defect region, the first determination process is performed in the determination step. When it is determined that the first estimated defect region is not the same as the second estimated defect region, the determination process is performed, In the first determination process, When at least one of the first estimated defect region and the second estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since at least one of the first estimated defect region and the second estimated defect region began to be detected, it is determined that the first estimated defect region and the second estimated defect region are defects. If at least one of the first estimated defect region and the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since at least one of the first estimated defect region and the second estimated defect region began to be detected, then it is determined that the first estimated defect region and the second estimated defect region are not defects. When both the first estimated defect region and the second estimated defect region are not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region and the second estimated defect region are not defects. In the second determination process described above, When the first estimated defect region is located at the predetermined position relative to the molten pool, and a predetermined time has elapsed since the first estimated defect region was first detected, it is determined that the first estimated defect region is a defect. If the first estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the first estimated defect region was detected, it is determined that the first estimated defect region is not a defect. When the first estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the first estimated defect region is not a defect. When the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is a defect. If the second estimated defect region is located at the predetermined position relative to the molten pool, and the predetermined time has not elapsed since the second estimated defect region was first detected, it is determined that the second estimated defect region is not a defect. The defect detection method according to claim 1, wherein when the second estimated defect region is not in the predetermined position relative to the molten pool, it is determined that the second estimated defect region is not a defect.

9. In the detection step, When the positional relationship between the first estimated defect region and the molten pool is the same as the positional relationship between the second estimated defect region and the molten pool, the first estimated defect region is estimated to be identical to the second estimated defect region. The defect detection method according to claim 8, wherein when the positional relationship between the first estimated defect region and the molten pool is different from the positional relationship between the second estimated defect region and the molten pool, it is estimated that the first estimated defect region is not the same as the second estimated defect region.

10. The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Based on the judgment result, an adjustment step is made to adjust the conditions of the arc welding, Equipped with, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. An arc welding method in which, when the estimated defect region is not located in the predetermined position relative to the molten pool, it is determined that the estimated defect region is not a defect.

11. A camera that acquires images of the area around the molten pool during arc welding, A control device that determines whether or not the estimated defect region in the aforementioned image is actually a defect, Equipped with, The control device is The estimated defect region and the molten pool are detected, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. A defect detection device that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

12. The camera includes a first camera and a second camera, The first camera is positioned in front of the molten pool in the direction of welding. The defect detection device according to claim 11, wherein the second camera is positioned behind the molten pool in the direction of welding.

13. The defect detection device according to claim 11, wherein the depression angle of the camera is 15° or more and 60° or less.

14. The defect detection device according to claim 11, wherein, in a plan view, the angle between the straight line connecting the camera and the molten pool and the straight line along the direction of welding is 0° or more and 60° or less.

15. A camera that acquires images of the area around the molten pool during arc welding, A control device that determines whether the estimated defect area in the aforementioned image is actually a defect, and adjusts the arc welding conditions based on the determination result, Torch and, Equipped with, The control device is The estimated defect region and the molten pool are detected, When the estimated defect region is located at a predetermined position relative to the molten pool, it is determined that the estimated defect region is a defect. An arc welding apparatus that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

16. On the computer, The acquisition process involves obtaining images of the area around the molten pool during arc welding, A detection step for detecting the estimated defect region and the molten pool in the aforementioned image, A determination step of determining whether or not the estimated defect region is a defect based on the positional relationship between the estimated defect region and the molten pool, Make it run, In the determination step, When the estimated defect region is located at a predetermined position relative to the molten pool, the estimated defect region is determined to be a defect. A program that determines that the estimated defect region is not a defect when the estimated defect region is not located in the predetermined position relative to the molten pool.

17. A recording medium on which the program described in claim 16 is recorded.

Images