Welding position identification system, welding position identification method, and welding position identification program

Abstract

The present invention comprises an imaging unit and a control unit, the imaging unit being configured to be capable of capturing an image of a region including at least a pair of adjacent metal pins, and the control unit being configured to be capable of executing pin position identification processing for identifying the position of each metal pin from image data captured by the imaging unit, and welding position identification processing for identifying the welding position of the pair of metal pins on the basis of the identified position of each metal pin.

Description

Welding position identification system, welding position identification method, and welding position identification program

[0001] The present invention relates to a welding position specifying system, a welding position specifying method, and a welding position specifying program.

[0002] Conventionally, there is a welding machine that includes a camera that photographs a product to be processed, an edge extraction unit, an invariant edge acquisition unit, a processing program correction unit, and a welding robot (see, for example, Patent Document 1). In the welding machine of Patent Document 1, the edge extraction unit extracts edge images from the captured image of the product photographed by the camera. Furthermore, the invariant edge acquisition unit acquires, as invariant edge images, edge images of invariant surfaces that do not shift in position relative to welding points set in the processing program.

[0003] The correction amount acquisition unit performs pattern matching between the master invariant edge image acquired by the invariant edge acquisition unit and the workpiece invariant edge image, and acquires the amount of deviation between them as a correction amount for the welding point. Furthermore, the machining program correction unit corrects the welding point by the correction amount and generates a corrected machining program for welding the workpiece. Then, the welding robot welds the workpiece based on the corrected machining program.

[0004] Patent No. 6286276

[0005] However, for example, when welding adjacent metal pins used in an electric motor or the like, the objects to be welded are small and the tolerance range for the welding position is limited, so there is a problem that even a slight deviation in the positioning of the welding robot's control program can result in a large welding error, and it is necessary to accurately analyze the positional relationship of the metal pins and determine the appropriate welding position.

[0006] One aspect of the present invention is a welding position identifying system, a welding position identifying method, and a welding position identifying program that can improve welding quality.

[0007] A welding position identification system according to one aspect of the present invention includes an imaging unit and a control unit, wherein the imaging unit is configured to be able to image an area including at least a pair of adjacent metal pins, and the control unit is configured to be able to execute a pin position identification process that identifies the position of each metal pin from image data captured by the imaging unit, and a welding position identification process that identifies the welding position of the pair of metal pins based on the identified positions of each metal pin.

[0008] A welding position identification method according to one aspect of the present invention includes an imaging process of imaging an area including at least a pair of adjacent metal pins using an imaging unit, a pin position identification process of identifying the positions of the metal pins from the captured image data, and a welding position identification process of identifying the welding position of the pair of metal pins.

[0009] A welding position identification program according to one aspect of the present invention causes a control device to execute an imaging process that causes an imaging unit to image an area that includes at least a pair of adjacent metal pins, a pin position identification process that identifies the positions of the metal pins from the image data that has been captured, and a welding position identification process that identifies the welding position of the pair of metal pins.

[0010] According to the welding position identification system, welding position identification method, and welding position identification program of one aspect of the present invention, the position of a metal pin is identified from image data captured by an imaging unit, and the welding position is identified based on the identified position, thereby improving welding quality.

[0011] FIG. 1 is a schematic diagram showing a welding position identifying system according to an embodiment of the present invention. FIG. 2 is a schematic diagram showing a welding jig according to the embodiment. FIG. 3 is a functional block diagram showing a control device according to the embodiment. FIG. 4 is a diagram showing an example of image data captured by a camera according to the embodiment. FIG. 5 is a schematic diagram showing a region including at least a pair of adjacent metal pins in the image data according to the embodiment. FIG. 6 is a schematic diagram showing a pin region identifying process according to the embodiment. FIG. 7 is a schematic diagram showing a center position candidate identifying process according to the embodiment. FIG. 8 is a schematic diagram showing a first center position identifying process according to the embodiment. FIG. 9 is a schematic diagram showing the first center position identifying process according to the embodiment. FIG. 10 is a schematic diagram showing the first center position identifying process according to the embodiment. FIG. 11 is a schematic diagram showing a second center position identifying process according to the embodiment. FIG. 12 is a schematic diagram showing a welding position identifying process according to the embodiment. FIG. 13 is a schematic diagram showing a welding position identifying process according to the embodiment. FIG. 14 is a schematic diagram showing a welding position identifying process according to the embodiment. FIG. 15 is a flowchart showing an example of a welding position identifying method according to the embodiment.

[0012] The best mode for carrying out the present invention will be described below with reference to the drawings. Note that the following embodiments do not limit the inventions according to the claims, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the invention.

[0013] [Overall Configuration of Welding Position Identifying System According to the Present Embodiment] Fig. 1 is a schematic diagram showing a welding position identifying system according to an embodiment of the present invention. First, with reference to Fig. 1, an overview of the welding position identifying system 1 according to the embodiment of the present invention will be described. As shown in Fig. 1, the welding position identifying system 1 according to the present embodiment generally includes a camera 50 that functions as an imaging unit, and a control device 100. The welding position identifying system 1 also includes a welding jig 10 and a welding head 30.

[0014] Fig. 2 is a schematic diagram showing a welding jig of this embodiment. As shown in Fig. 2, welding jig 10 is configured to be able to attach a plurality of metal pins P (flat wires) to be welded. Furthermore, welding jig 10 has a rotating portion and is configured to be rotatable. Note that, since various known configurations can be adopted for welding jig 10, detailed description thereof will be omitted.

[0015] 1, the welding head 30 is disposed above the welding jig 10 and is configured to be able to weld a pair of adjacent metal pins P. A laser beam oscillated by a laser oscillator (not shown) is sent to the welding head 30 via an optical fiber. The welding head 30 welds the pair of adjacent metal pins P by irradiating the metal pins P with laser light from a nozzle (not shown).

[0016] The welding head 30 is attached to a Cartesian coordinate moving device or the like and is configured to be movable in three axial directions, i.e., forward / backward, left / right, and depth directions, on the welding jig 10. The welding head 30 also has a tilt axis and a rotation axis, allowing the direction of laser light irradiation to be freely changed. Note that the welding head 30 can employ various known configurations, and therefore detailed description thereof will be omitted.

[0017] Fig. 4 is a diagram showing an example of image data captured by the camera of this embodiment. The camera 50 is provided with one lens and one image sensor, and is disposed above the welding jig 10 so as to be able to capture an image of an area including at least a pair of adjacent metal pins P, as shown in Figs. 1 and 4. In this embodiment, the camera 50 is configured to capture an image of the plurality of metal pins P attached to the welding jig 10 from above, as shown in Fig. 1.

[0018] Camera 50 is configured to supply captured image data 122 to control device 100. Note that camera 50 according to this embodiment may be configured to output a digital signal (image data 122) directly from camera 50 to control device 100, or may be configured to convert an analog signal (captured image signal) output from camera 50 into a digital signal (image data 122) using an A / D converter (not shown) or the like and output the digital signal to control device 100.

[0019] Fig. 3 is a functional block diagram showing a control device of this embodiment. As shown in Fig. 3, the control device 100 includes a control unit 110 and a storage unit 120. The control device 100 according to this embodiment is, for example, a numerical control device or an electronic computer such as a desktop personal computer, a laptop computer, or a tablet terminal.

[0020] Control device 100 is configured to be able to control camera 50. Control device 100 also acquires image data 122 captured by camera 50 and stores the image data in storage unit 120. Control device 100 may also be configured to be able to control welding head 30.

[0021] The control unit 110 is configured by, for example, an integrated processing device having a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit). As shown in FIG. 3 , the control unit 110 includes a pin position specifying unit 112, a center position specifying unit 114, and a welding position specifying unit 116.

[0022] The control unit 110 is configured to be able to execute a pin position specifying process that specifies the position of each metal pin P from the image data 122 captured by the camera 50, and a welding position specifying process that specifies the welding position W of a pair of metal pins P based on the specified position of each metal pin P. The pin position specifying process includes a pin position candidate specifying process that specifies candidates for the position of the metal pin P from the image data 122, and a center position specifying process that specifies the center position C of each metal pin P.

[0023] In this embodiment, the center position C of the metal pin P generally refers to the centroid position of the metal pin P when the metal pin P attached to the welding jig 10 is viewed from above in a plan view, but is not limited to this. For example, if the tip of the metal pin P is formed in a convex shape, the apex of the tip (the highest point of the convex shape) may be identified as the center position C.

[0024] Fig. 5 is a schematic diagram showing an area including at least a pair of adjacent metal pins in the image data of this embodiment. Fig. 6 is a schematic diagram showing the pin area identification process of this embodiment. The pin position identification unit 112 is configured to be able to execute the pin position candidate identification process. As shown in Figs. 5 and 6, the pin position candidate identification process includes a pin area identification process that identifies a rectangular area A including a pair of metal pins P from the image data 122 by color detection.

[0025] For ease of explanation, in this embodiment, the image data 122 will be described as image data that shows only some of the metal pins P attached to the welding jig 10, but this is not limited to this, and the image data 122 may be image data that shows all of the metal pins P, as shown in FIG. 4.

[0026] In the pin region identification process, the pin position identification unit 112 identifies the regions of the pair of metal pins P by color detection in the HSV color space. Specifically, the pin position identification unit 112 identifies the regions of the pair of metal pins P by color detection based on the difference between the color of the surface of the welding jig 10 and the color of the surface of the metal pin P. Furthermore, if there is a blown-out highlight region in the image data 122, the pin position identification unit 112 can interpolate the region from surrounding image regions.

[0027] 7 is a schematic diagram showing the center position candidate specifying process of this embodiment. The pin position candidate specifying process includes a center position candidate specifying process for specifying center position candidates CC of each metal pin P from the image data 122 by machine learning, as shown in FIG. 7. In the pin position candidate specifying process, the pin position specifying unit 112 analyzes the image data 122 based on the teacher data 124 stored in the storage unit 120, and specifies the center position candidates CC of each metal pin P.

[0028] In this embodiment, the teacher data 124 is created such that an image including only the metal pin P is defined as correct data (positive image) and an image not including the metal pin P is defined as incorrect data (negative image), and the pin position identification unit 112 is configured to perform analysis using a cascade detector.

[0029] 8 is a schematic diagram showing the first center position identifying process of this embodiment. The center position identifying unit 114 is configured to be able to execute the center position identifying process. In this embodiment, the center position identifying unit 114 is configured to be able to execute the first center position identifying process using image processing and the second center position identifying process using machine learning. In the first center position identifying process, the center position identifying unit 114 identifies the center position C of each metal pin P based on the luminance change in at least two directions of the rectangular area A identified in the pin area identifying process, as shown in FIG.

[0030] 9 and 10 are schematic diagrams illustrating the first center position identification process of this embodiment. Specifically, the center position identification unit 114 identifies a candidate CC for the center position of each metal pin P based on the change in luminance in the direction of the long side of the rectangular region A, as shown in Fig. 8 and Fig. 9, and identifies the center position C of each metal pin P based on the change in luminance in the direction of the short side passing through the candidate CC for the center position, as shown in Fig. 10.

[0031] 11 is a schematic diagram showing the second center position identification process of this embodiment. In the second center position identification process, the center position identification unit 114 identifies the center position C of each metal pin P based on the color distribution around the center position candidate CC. Specifically, as shown in FIG. 11 , the center position identification unit 114 searches for an area where the same color or similar colors are distributed in large numbers around the center position candidate CC of each metal pin P identified by the pin position identification unit 112 from the image data 122 by machine learning in the center position candidate identification process, and identifies the area as the center position C of the metal pin P.

[0032] In addition, the center position identification unit 114 is configured to be able to execute a candidate identification determination process that determines whether a candidate center position CC has been identified through machine learning, and can selectively use the first center position identification process and the second center position identification process based on the determination result.

[0033] Specifically, if the pin position identifying unit 112 has not identified a center position candidate CC, the center position identifying unit 114 executes a first center position identifying process to identify the center position C of each metal pin P based on the luminance change within the rectangular area A identified by the pin position identifying unit 112 in the pin area identifying process. Furthermore, if the pin position identifying unit 112 has identified a center position candidate CC, the center position identifying unit 114 executes a second center position identifying process to identify the center position C of each metal pin P based on the color distribution around the center position candidate CC.

[0034] 12 to 14 are schematic diagrams showing the welding position identification process of this embodiment. The welding position identification unit 116 is configured to be able to execute the welding position identification process. In the welding position identification process, the welding position identification unit 116 assumes the intermediate position between the pair of metal pins P as a candidate welding position, as shown in FIG. 12 . Then, the welding position identification unit 116 extracts the luminance distribution on a line passing through each center position C of the pair of metal pins P and the candidate welding position, and identifies the position on the line between the metal pins P with the lowest luminance value as the welding position W, as shown in FIGS. 13 and 14 .

[0035] Furthermore, as a preprocessing step for the welding position identification process, the welding position identifying unit 116 may perform a smoothing process on the luminance distribution on a line that passes through each of the center positions C of the pair of metal pins P and the candidate welding positions. By performing the smoothing process, if there is a position on the line with an extremely low luminance value, it is possible to prevent the position from being erroneously identified as the welding position W.

[0036] The control unit 110 having the above configuration is configured to be able to create a control program for the welding head 30 for welding the metal pin P, based on the identified welding position W. The control unit 110 is also configured to be able to correct the control program for the welding head 30, based on the identified welding position W.

[0037] The storage unit 120 has a storage medium such as a hard disk drive (HDD) or a solid state drive (SSD), and stores various data in a readable and writable manner. As shown in Fig. 2, the storage unit 120 stores image data 122, training data 124, and a welding position identification program 126. Furthermore, the storage unit 120 stores programs necessary for controlling each part of the control device 100.

[0038] The welding position specifying program 126 causes the control device 100 to execute an imaging process for causing the imaging unit (camera 50 in this embodiment) to capture an image of an area including at least a pair of adjacent metal pins P, a pin position specifying process for specifying the positions of the metal pins P from the captured image data 122, and a welding position specifying process for specifying the welding position W of the pair of metal pins P. The welding position specifying program 126 also includes information on the number of metal pins P attached to the welding jig 10. Including the information on the number of metal pins P attached to the welding jig 10 has the advantage of preventing omission of detection of metal pins P included in the image data 122.

[0039] [Welding position identifying method according to the present embodiment] Next, a description will be given of the welding position identifying method of the welding position identifying system 1 according to the present embodiment. The welding position identifying method of the welding position identifying system 1 according to the present embodiment generally includes an imaging step of imaging an area including at least a pair of adjacent metal pins P using an imaging unit (camera 50 in the present embodiment), a pin position identifying step of identifying the positions of the metal pins P from the captured image data 122, and a welding position identifying step of identifying the welding position W of the pair of metal pins P.

[0040] 15 is a flowchart showing an example of a welding position identification method according to this embodiment. First, the camera 50 captures an image of an area including at least a pair of adjacent metal pins P (S1 in FIG. 15: photographing step). Specifically, the camera 50 captures an image of a plurality of metal pins P attached to the welding jig 10. Then, the control unit 110 of the control device 100 acquires image data 122 captured by the camera 50.

[0041] Next, the control unit 110 of the control device 100 identifies the position of each metal pin P from the image data 122 captured by the camera 50 (S10 in FIG. 15: pin position identification step). The pin position identification step will be described in detail below. First, the pin position identification unit 112 of the control unit 110 identifies candidate positions of the metal pins P from the image data 122 (S11 in FIG. 15: pin position candidate identification step).

[0042] In the pin position candidate identifying step, the pin position identifying unit 112 identifies a rectangular area A including the pair of metal pins P by color detection from the image data 122 (S111 in FIG. 15 : pin area identifying step). Specifically, the pin position identifying unit 112 identifies the area of ​​the pair of metal pins P based on the difference in color between the surface of the welding jig 10 and the surface of the metal pins P by color detection in the HSV color space.

[0043] Furthermore, in a pin position candidate identifying step, the pin position identifying unit 112 identifies candidates CC for the center positions of each metal pin P from the image data 122 by machine learning (S112 in FIG. 15 : center position candidate identifying step). Specifically, the pin position identifying unit 112 identifies candidates CC for the center positions of each metal pin P from the image data 122 using a cascade classifier based on the teacher data 124 in the storage unit 120.

[0044] Next, the center position specifying unit 114 specifies the center position C of each metal pin P (center position specifying step). Specifically, the center position specifying unit 114 determines whether or not the pin position specifying unit 112 has specified a center position candidate CC by machine learning in the center position candidate specifying step (S12 in FIG. 15 : candidate specification determining step).

[0045] When the pin position specifying unit 112 specifies a center position candidate CC (YES in S12 of FIG. 15), the center position specifying unit 114 specifies the center position C based on the color distribution around the specified center position candidate CC (S13 of FIG. 15: second center position specifying step). Specifically, the center position specifying unit 114 detects an area around the center position candidate CC where the same color or similar colors are distributed, and specifies the area as the center position C of the metal pin P.

[0046] On the other hand, if the pin position specifying unit 112 has not specified a center position candidate CC (NO in S12 in FIG. 15 ), the center position specifying unit 114 specifies the center position C of each metal pin P based on the luminance change within the rectangular area A specified by the pin position specifying unit 112 in the pin area specifying step (S14 in FIG. 15 : first center position specifying step). Specifically, the center position specifying unit 114 specifies the center position candidate CC of each metal pin P based on the luminance change in the long side direction of the specified rectangular area A, and specifies the center position C of each metal pin P based on the luminance change in the short side direction passing through the center position candidate CC.

[0047] Thereafter, the welding position identifying unit 116 identifies the welding position W of the pair of metal pins P based on the positions of each metal pin P identified in the pin position identifying process (S20 in FIG. 15 : welding position identifying step). Specifically, the welding position identifying unit 116 assumes the midpoint between the pair of metal pins P as a candidate welding position, extracts the luminance distribution on a line passing through each center position C of the pair of metal pins P and the candidate welding position, and identifies the position on the line with the lowest luminance value between each metal pin P as the welding position W. Through the above steps, a series of welding position identifying methods are performed by the welding position identifying system 1 according to this embodiment.

[0048] [Advantages of the welding position identifying system, welding position identifying method, and welding position identifying program according to the present embodiment] As described above, the welding position identifying system 1 according to the present embodiment includes an imaging unit (camera 50 in the present embodiment) and a control unit 110. The imaging unit is configured to be able to image an area that includes at least a pair of adjacent metal pins P, and the control unit 110 is configured to be able to execute a pin position identifying process that identifies the position of each metal pin P from image data 122 captured by the imaging unit, and a welding position identifying process that identifies the welding position W of the pair of metal pins P based on the identified position of each metal pin P.

[0049] The welding position identifying system 1 according to this embodiment has such a configuration, and thus identifies the position of the metal pin P from the image data 122 captured by the imaging unit, and identifies the welding position W of the pair of metal pins P based on the identified position of the metal pin P, thereby having the advantage of being able to improve welding quality.

[0050] Furthermore, in the welding position identifying system 1 according to this embodiment, the pin position identifying process includes a pin position candidate identifying process that identifies candidates for the positions of the metal pins P from the image data 122, and a center position identifying process that identifies the center positions C of the metal pins P. Having such a configuration has the advantage of reducing the analysis process of the welding positions W (processing required to identify the welding positions W).

[0051] Furthermore, in the welding position identifying system 1 according to this embodiment, the pin position candidate identifying process includes a pin area identifying process that identifies a rectangular area A including a pair of metal pins P by color detection from the image data 122. This configuration has the advantage of being able to reduce false detections compared to detecting the shape of the metal pins P by edge extraction, for example.

[0052] Furthermore, in the welding position identifying system 1 according to this embodiment, in the pin area identifying process, the control unit 110 identifies the areas of the pair of metal pins P by color detection in the HSV color space. With this configuration, the areas of the pair of metal pins P are identified taking into consideration differences in saturation and brightness, which has the advantage of enabling more accurate area identification and ultimately improving welding quality.

[0053] Furthermore, in the welding position identifying system 1 according to this embodiment, the pin position candidate identifying process includes a center position candidate identifying process that identifies candidates CC for the center position of each metal pin P by machine learning from the image data 122. By providing such a configuration, there is an advantage that the center position candidates CC can be identified in a short time.

[0054] Furthermore, in the welding position identifying system 1 according to this embodiment, the control unit 110 identifies the center position C of each metal pin P in the center position identifying process based on the luminance changes in at least two directions of the rectangular area A identified in the pin area identifying process. This configuration has the advantage of being able to identify the center position C without identifying the center position candidate CC.

[0055] Furthermore, in the welding position identifying system 1 according to this embodiment, the control unit 110 identifies a candidate center position CC of each metal pin P based on the change in brightness in the long side direction of the rectangular area A, and identifies the center position C of each metal pin P based on the change in brightness in the short side direction passing through the candidate center position CC. This configuration has the advantage of being able to identify the center position C efficiently and accurately, thereby improving welding quality.

[0056] Furthermore, in the welding position identifying system 1 according to this embodiment, the control unit 110 identifies the center position C of each metal pin P based on the color distribution around the center position candidate CC in the center position identifying process. This configuration has the advantage of being able to identify the center position C more quickly and accurately than identification based on brightness changes, thereby further improving welding quality.

[0057] Furthermore, in the welding position identifying system 1 according to this embodiment, the pin position candidate identifying process includes a pin area identifying process that identifies a rectangular area A containing a pair of metal pins P by color detection from the image data 122, and the control unit 110 is configured to execute a candidate identifying determination process that determines whether a center position candidate CC has been identified by machine learning. If a center position candidate CC has not been identified, the center position identifying process identifies the center position C of each metal pin P based on the luminance change within the rectangular area A identified in the pin area identifying process. If a center position candidate CC has been identified, the center position identifying process identifies the center position C based on the color distribution around the center position candidate CC. This configuration has the advantage of automatically selecting a more accurate second center position identifying process when a center position candidate CC is present, thereby further improving welding quality.

[0058] Furthermore, in the welding position identifying system 1 according to this embodiment, the control unit 110 assumes the midpoint between the pair of metal pins P as a candidate welding position in the welding position identifying process, extracts the luminance distribution on a line passing through each center position C of the pair of metal pins P and the candidate welding position, and identifies the position on the line with the lowest luminance value between the metal pins P as the welding position W. This configuration has the advantage of enabling accurate identification of the welding position, thereby improving welding quality.

[0059] [Modifications] While the preferred embodiments of the present invention have been described above, the technical scope of the present invention is not limited to the scope described in the above-described embodiments. Various modifications and improvements can be made to the above-described embodiments.

[0060] For example, in the above-described embodiment, the pin position identification process has been described as including a pin position candidate identification process that identifies candidates for the positions of the metal pins P from the image data 122 and a center position identification process that identifies the center positions C of each metal pin P, but is not limited to this. The pin position identification process does not have to include both the pin position candidate identification process and the center position identification process, or may include only one of them.

[0061] In the above-described embodiment, the pin position candidate identification process has been described as including a pin area identification process that identifies a rectangular area A containing a pair of metal pins P by color detection from the image data 122, but is not limited to this, and the pin position candidate identification process does not need to include the pin area identification process.

[0062] In the above-described embodiment, the control unit 110 is described as identifying the region of the pair of metal pins P by color detection in the HSV color space in the pin region identification process, but this is not limiting. In the pin region identification process, the control unit 110 does not have to identify the region of the pair of metal pins P by color detection in the HSV color space. For example, the control unit 110 may identify the region of the pair of metal pins P by color detection in the RGB color space.

[0063] In the above-described embodiment, the pin position candidate identification process has been described as including a center position candidate identification process that identifies candidates CC for the center position of each metal pin P from the image data 122 by machine learning, but is not limited to this, and the pin position candidate identification process does not necessarily have to include a center position candidate identification process.

[0064] In the above-described embodiment, the control unit 110 is described as specifying the center position C of each metal pin P in the center position specifying process based on the luminance changes in at least two directions of the rectangular area A specified in the pin area specifying process, but this is not limited to this. In the center position specifying process, the control unit 110 does not have to specify the center position C of each metal pin P based on the luminance changes in at least two directions of the rectangular area A specified in the pin area specifying process. For example, the control unit 110 may specify the center position C of each metal pin P based on the luminance change in one direction of the rectangular area A.

[0065] In the above-described embodiment, the control unit 110 identifies the candidate center position CC of each metal pin P based on the change in brightness in the long side direction of the rectangular area A, and identifies the center position C of each metal pin based on the change in brightness in the short side direction passing through the candidate center position CC. However, this is not limited to this. The control unit 110 does not have to identify the candidate center position CC of each metal pin P based on the change in brightness in the long side direction of the rectangular area A and identify the center position C of each metal pin based on the change in brightness in the short side direction passing through the candidate center position CC. For example, the control unit 110 may identify the candidate center position CC of each metal pin P based on the change in brightness in the short side direction of the rectangular area A, and identify the center position C of each metal pin based on the change in brightness in the long side direction passing through the candidate center position CC. Alternatively, the control unit 110 may identify the center position C of each metal pin based on the change in brightness in the diagonal direction of the rectangular area A.

[0066] In the above-described embodiment, the control unit 110 is described as specifying the center position C of each metal pin P based on the color distribution around the center position candidate CC in the center position specifying process, but this is not limiting. In the center position specifying process, the control unit 110 does not have to specify the center position C of each metal pin P based on the color distribution around the center position candidate CC.

[0067] In the above-described embodiment, the pin position candidate identification process includes a pin area identification process that identifies a rectangular area A containing a pair of metal pins P by color detection from the image data 122, and the control unit 110 is configured to be able to execute a candidate identification determination process that determines whether a center position candidate CC has been identified by machine learning. If a center position candidate CC has not been identified, the center position identification process identifies the center position C of each metal pin P based on the luminance change within the rectangular area A identified in the pin area identification process. If a center position candidate CC has been identified, the center position identification process identifies the center position C based on the color distribution around the center position candidate CC. However, this is not limiting. The control unit 110 does not necessarily have to be able to execute the candidate identification determination process.

[0068] In the above-described embodiment, the control unit 110 assumes the intermediate position of the pair of metal pins P as a candidate welding position in the welding position identification process, extracts the luminance distribution on a line that passes through the center positions C of the pair of metal pins P and the candidate welding position, and identifies the position on the line where the luminance value is lowest between the metal pins P as the welding position W. However, the present invention is not limited to this. In the welding position identification process, the control unit 110 does not have to assume the intermediate position of the pair of metal pins P as a candidate welding position, extracts the luminance distribution on a line that passes through the center positions C of the pair of metal pins P and the candidate welding position, and identify the position on the line where the luminance value is lowest between the metal pins P as the welding position W.

[0069] In the above-described embodiment, camera 50 has been described as being attached above welding jig 10 on a ceiling surface or the like, but this is not limited thereto. Camera 50 may be built into welding head 30 and pointed in the same direction as the direction in which welding head 30 emits laser light. Camera 50 may also be attached to the outside of welding head 30. When camera 50 is built into welding head 30 or attached to the outside of welding head 30, control device 100 may rotate image data 122 to match the attitude of welding head 30 (camera 50).

[0070] REFERENCE SIGNS LIST 1 Welding position identification system 10 Welding jig 30 Welding head 50 Camera 100 Control device 110 Control unit 112 Pin position identification unit 114 Center position identification unit 116 Welding position identification unit 120 Storage unit 122 Image data 124 Training data 126 Welding position identification program A Rectangular area C Center position CC Candidate for center position P Metal pin W Welding position

Claims

1. A welding position identification system comprising an imaging unit and a control unit, wherein the imaging unit is configured to be able to image an area including at least a pair of adjacent metal pins, and the control unit is configured to be able to execute a pin position identification process for identifying the position of each metal pin from image data captured by the imaging unit, and a welding position identification process for identifying the welding position of the pair of metal pins based on the identified positions of each metal pin.

2. The welding position identifying system according to claim 1, wherein the pin position identifying process includes a pin position candidate identifying process that identifies candidates for the positions of the metal pins from the image data, and a center position identifying process that identifies the center positions of the metal pins.

3. The welding position identifying system according to claim 2, wherein the pin position candidate identifying process includes a pin area identifying process for identifying a rectangular area including the pair of metal pins by color detection from the image data.

4. The welding position identifying system according to claim 3, wherein in the pin area identifying process, the control unit identifies the areas of the pair of metal pins by color detection in an HSV color space.

5. A welding position identifying system according to any one of claims 2 to 4, wherein the pin position candidate identifying process includes a center position candidate identifying process that identifies candidates for the center position of each metal pin from the image data by machine learning.

6. The welding position identification system according to claim 3 or 4, wherein in the center position identification process, the control unit identifies the center position of each metal pin based on a change in brightness in at least two directions of the rectangular area identified in the pin area identification process.

7. The welding position identification system according to claim 6, wherein the control unit identifies candidates for the center position of each metal pin based on the change in luminance in the direction of the long side of the rectangular area, and identifies the center position of each metal pin based on the change in luminance in the direction of the short side passing through the candidates for the center position.

8. The welding position identifying system according to claim 5, wherein in the center position identifying process, the control unit identifies the center position of each metal pin based on a color distribution around the center position candidate.

9. The welding position identification system according to claim 5, wherein the pin position candidate identification process includes a pin area identification process for identifying a rectangular area including the pair of metal pins by color detection from the image data, and the control unit is configured to be capable of executing a candidate identification determination process for determining whether or not the center position candidate has been identified by the machine learning, and if the center position candidate has not been identified, the center position identification process identifies the center position of each metal pin based on a luminance change within the rectangular area identified in the pin area identification process, and if the center position candidate has been identified, the center position identification process identifies the center position based on a color distribution around the center position candidate.

10. The welding position identification system according to claim 1 or 2, wherein in the welding position identification process, the control unit assumes the midpoint of the pair of metal pins as a candidate for the welding position, extracts a luminance distribution on a straight line passing through each center position of the pair of metal pins and the candidate for the welding position, and identifies the position on the straight line between the metal pins with the lowest luminance value as the welding position.

11. A welding position identifying method comprising: an imaging step of imaging an area including at least a pair of adjacent metal pins with an imaging unit; a pin position identifying step of identifying the positions of the metal pins from the captured image data; and a welding position identifying step of identifying the welding position of the pair of metal pins.

12. A welding position identification program that causes a control device to execute an imaging process that causes an imaging unit to image an area that includes at least a pair of adjacent metal pins, a pin position identification process that identifies the positions of the metal pins from the captured image data, and a welding position identification process that identifies the welding position of the pair of metal pins.

Images