JP2023098300A - Position detection device, positioning device, welding device, position detection program, and position detection method - Google Patents

Abstract

To provide a position detection device that can maintain the accuracy of position detection for a target with low shape reproducibility.SOLUTION: In the position detection device, a control unit 104 includes: an image storage unit 115 that stores a registered image of each of a plurality of first time sections based on a photographed image of a target in each of the first time sections; a pattern matching unit 112 that performs pattern matching of the registered image of each of the first time sections and a photographed image of the target obtained in a second time section different from the plurality of first time sections to obtain a matching position of the target for each registered image; and a position determination unit (consultation unit 113) that determines a detected position of the target based on the matching position of each registered image.SELECTED DRAWING: Figure 2

Description

The present invention relates to a position detection device, a positioning device, a welding device, a position detection program, and a position detection method.

Conventionally, there is a technique of position detection by pattern matching.

For example, in Patent Document 1, in a fingerprint recognition device, an input fingerprint image is sequentially compared with dictionary images registered in a plurality of dictionaries. There is proposed a technique for updating one dictionary with a dictionary image registered in one dictionary and updating the other dictionary with a dictionary image extracted from an input fingerprint image.

JP-A-7-271981

However, in the case of an object with low shape reproducibility, even if pattern matching is successful, it may not be suitable for registration as a new dictionary image.
SUMMARY OF THE INVENTION Accordingly, it is an object of the present invention to maintain the accuracy of position detection even for an object whose shape has low reproducibility.

One aspect of the position detection device according to the present invention is an image storage unit that stores registered images for each of the first time intervals based on captured images of an object in each of a plurality of first time intervals; pattern matching is performed between each registered image and the photographed images of the object obtained in a second time interval different from the plurality of first time intervals, and matching positions of the object are determined for each of the registered images. and a position determination unit that determines the detection position of the object based on the matching position for each of the registered images.

Further, one aspect of a positioning device according to the present invention includes the position detection device and a positioning mechanism that positions an object to be positioned using the detection position obtained by the position detection device as a reference.

Further, one aspect of the welding apparatus according to the present invention includes the position detection device, a positioning mechanism for positioning a welding point with reference to the detection position obtained by the position detection device, and a weld positioned by the positioning mechanism. and a welding part for welding the part.
Further, one aspect of the position detection program according to the present invention causes a computer to function as each component in the position detection device.

Further, one aspect of the position detection method according to the present invention is a registered image for each first time interval based on a photographed image of an object in each of the plurality of first time intervals, and the plurality of first time intervals. a matching process of performing pattern matching with photographed images of the object obtained in different second time intervals to obtain a matching position of the object for each registered image; and a position determination step of determining the detected position of the object.

According to the present invention, the accuracy of position detection can be maintained even for an object whose shape has low reproducibility.

FIG. 1 is a diagram schematically showing one embodiment of the welding device of the present invention. FIG. 2 is a functional block diagram showing the functional structure of the controller. FIG. 3 is a flow chart showing the processing procedure of position control in the control unit. FIG. 4 is a diagram showing an example of a photographed image obtained by a camera. FIG. 5 is a diagram showing an example of a registered image stored in an image storage unit. FIG. 6 is a diagram showing an example of determination of detection positions by consensus. FIG. 7 is a flow chart showing a processing procedure for updating a registered image by the image registration unit.

Hereinafter, embodiments of a position detection device, a positioning device, a welding device, a position detection program, and a position detection method of the present disclosure will be described in detail with reference to the accompanying drawings. However, in order to avoid unnecessary redundancy in the following description and facilitate the understanding of those skilled in the art, more than necessary detailed description may be omitted. For example, detailed descriptions of well-known matters and redundant descriptions of substantially the same configurations may be omitted. In addition, the elements described in the previously described figures may be appropriately referred to in the description of later figures.

<Structure of Welding Device>
FIG. 1 is a diagram schematically showing one embodiment of the welding device of the present invention.
The welding device 100 of this embodiment is a laser welder, and irradiates the tip of the copper wire 301 of the workpiece 300 with a laser beam 105 to weld the workpiece 300 .

The copper wire 301 of the workpiece 300 is cut by the blade 200 before welding, and the cut tip portion is irradiated with the laser beam 105 . The shape of the cut portion of the copper wire 301 is similar to some extent in any workpiece 300 . However, the shape of the cut portion fluctuates, for example, within a day, and also changes over time on a time scale of several days to several weeks. That is, the shape of the cut portion of the copper wire 301 has low reproducibility.

Welding device 100 includes camera 101 , welding section 102 , positioning arm 103 , and control section 104 .

A camera 101 photographs the workpiece 300 and outputs image data.
The welding part 102 emits a laser beam 105 for welding.
Positioning arm 103 positions welded portion 102 with respect to workpiece 300 .
The camera 101, the welding part 102 and the positioning arm 103 are connected to the control part 104 by wire or wirelessly.

Control unit 104 detects the position of copper wire 301 based on image data obtained by camera 101, and controls positioning arm 103 to position welded portion 102 at the detected position.

The control unit 104 corresponds to one embodiment of the position detection device of the present invention. In this embodiment, the control unit 104 is implemented by, for example, a personal computer, and the personal computer incorporates an embodiment of the position detection program of the present invention.

A combination of the control unit 104 and the positioning arm 103 corresponds to one embodiment of the positioning device of the present invention. The positioning arm 103 corresponds to an example of the positioning mechanism according to the present invention, and positions an object to be positioned (for example, the welded portion 102) on the basis of the detected position obtained by the control section 104.

FIG. 2 is a functional block diagram showing the functional structure of the control section 104. As shown in FIG. The functional block diagram shown in FIG. 2 also represents the program structure in one embodiment of the location detection program of the present invention. One embodiment of the position detection program causes a computer to function as each component corresponding to each functional block shown in FIG.

The control unit 104 includes an image acquisition unit 111 , a pattern matching unit 112 , a discussion unit 113 , a position control unit 114 , an image storage unit 115 and an image registration unit 116 .
An image acquisition unit 111 acquires image data as a photographed image of the workpiece 300 from the camera 101 .

The pattern matching unit 112 performs pattern matching between the captured image obtained by the camera 101 and the registered image stored in the image storage unit 115 to obtain matching positions. As the registered image, an image of the copper wire 301 in the captured image of the workpiece 300 captured by the camera 101 in the past is stored. The image storage unit 115 stores a plurality of registered images, and the pattern matching unit 112 performs pattern matching using each registered image to obtain each matching position. Note that the pattern matching section is sometimes simply referred to as a matching section.

The discussion unit 113 decides the detection position of the copper wire 301 through discussion based on the matching positions of each of the plurality of registered images.
The image registration unit 116 updates the registered image stored in the image storage unit 115 as necessary, and stores a new image of the copper wire 301 captured by the camera 101 in the image storage unit 115 as a registered image. Image registration unit 116 automatically updates the registered image, and when receiving a reset instruction from the user of welding apparatus 100, updates the registered image to a greater extent than the automatic update.

<Position control processing procedure>
FIG. 3 is a flowchart showing the processing procedure of position control in the control unit 104. As shown in FIG. FIG. 4 is a diagram showing an example of a photographed image obtained by the camera 101. As shown in FIG. FIG. 5 is a diagram showing an example of a registered image stored in the image storage unit 115. As shown in FIG. FIG. 6 is a diagram showing an example of determination of detection positions by consensus. Hereinafter, the processing procedure shown in the flowchart of FIG. 3 will be described with reference to FIGS. 4 to 6 as appropriate.

The position control process shown in the flowchart of FIG. 3 is executed each time the workpiece 300 is welded.
When the position control process is started, first, in step S<b>101 , a photographed image is acquired from the camera 101 by the image acquiring unit 111 . As shown in FIG. 4, the captured image 310 includes the copper wire 301 and the caulked portion 305, which is not shown in FIG. The copper wire 301 is welded to the crimped portion 305 by irradiating the copper wire 301 with the laser beam 105 . Therefore, it is required to accurately detect the position of the copper wire 301 in the captured image 310 including the caulked portion 305 .

The image portion of the copper wire 301 includes a cutting edge 302 and two cut surfaces 303 with the cutting edge 302 therebetween. The position, inclination, curvature, and the like of the cutting edge 302 and the cutting surface 303 fluctuate each time the copper wire 301 is cut, and change over time as the blade 200 wears. As a result, the images of the cut edge 302 and the cut surface 303 also fluctuate and change over time.
Accurate position detection by pattern matching is required for such an image of the copper wire 301 with low reproducibility.

In step S102, the pattern matching unit 112 performs pattern matching on the photographed image 310 obtained in step S101 of FIG. The pattern matching unit 112 performs pattern matching on each of the plurality of registered images stored in the image storage unit 115, as described above.

As shown in FIG. 5, multiple registered images 121 are registered in the image storage unit 115 . FIG. 5 shows six registered images 121 from a to f as an example. Each registered image 121 has a common basic shape, but differs in patterns such as reflection of the cut surface 303 . In step S102 of FIG. 3, such a plurality of registered images 121 are used for pattern matching, and a matching position is obtained for each registered image 121. FIG. Note that the process of finding a matching position for each registered image 121 is sometimes called a "matching process".

As each registered image 121, for example, an average image of a plurality of photographed images 310 obtained by photographing each workpiece 300 welded in one day by the camera 101 is used. That is, the image storage unit 115 stores the registered image 121 for each first time interval based on the photographed image 310 of the object (for example, the work 300) in each of a plurality of first time intervals (for example, each one day). Further, in this embodiment, the registered image 121 is an average image of a plurality of shot images obtained within the first time interval.

However, an image obtained by cutting out only the portion corresponding to the copper wire 301 in the average image is used as the registered image 121 . When the photographed images 310 are averaged, the photographed images 310 are aligned with reference to the detection positions detected when the workpieces 300 are welded.

Since the average image is an image in which fluctuations in the shape of the copper wire 301 are canceled, pattern matching with the image of the copper wire 301 in each workpiece 300 is highly likely to succeed, and is suitable as the registered image 121 .

As a result of pattern matching by the pattern matching unit 112, a matching position 131 corresponding to each registered image 121 is obtained as shown in FIG. That is, the pattern matching unit 112 combines the registered image 121 for each first time interval with the target object (for example, the work 300) obtained in a second time interval (for example, within another day) different from the plurality of first time intervals. pattern matching with the photographed image 310 is performed to obtain the matching position 131 of the workpiece 300 for each registered image 121 .

In the example shown in FIG. 6, some matching positions 131 (eg, matching positions 131 in registered images a, b, e, f) cluster at copper wire 301, while other matching positions 131 (eg, registered images c, d The matching positions 131 ) of the are separated from each other at different positions than the copper wire 301 . A matching position 131 deviated from the position of the copper wire 301 may also be obtained by using the registered image 121 of a plurality of patterns.

If the matching position 131 is deviated from the position of the copper wire 301, it is considered that the image of the crimped portion 305 and the registered image 121 coincidentally match each other to some extent. Conversely, when a plurality of matching positions 131 are gathered close to each other, it is considered that each of the plurality of registered images 121 matches the image of the copper wire 301 to a certain degree or more. It is thought that

Therefore, the accurate detection position of the copper wire 301 is determined by discussion in the discussion section 113 based on the matching positions 131 of each of the plurality of registered images 121 . The discussion section 113 corresponds to an example of the position determination section according to the present invention, and determines the detection position of the target object (for example, the work 300) based on the matching position 131 for each registered image. A process of determining the detected position of the object based on the matching position is sometimes called a "position determination process".

Even if the shape of the object is not reproducible, there is a certain degree of similarity between the captured images 310 . high position detection is realized.

Specifically, clustering is performed on a plurality of matching positions 131 by the discussion unit 113 in step S103 of FIG. That is, a cluster 132 is formed as shown in FIG. 6 for the matching positions 131 that are within a predetermined distance from each other. A matching position 131 that does not have another matching position 131 within a predetermined distance becomes a single cluster.

As a result of clustering, for the cluster 132 containing the maximum number of matching positions 131, the discussion unit 113 calculates the average position of the matching positions 131 belonging to the cluster 132 in step S104 of FIG. The control unit 104 uses this average position as the detection position of the copper wire 301 .

That is, the discussion unit 113 determines the average position 133 of the matching positions 131 in the cluster 132 to which the largest number of matching positions belong among the clusters 132 into which the matching positions 131 for each registered image are divided, as the detection position. By using the average of the most matching positions as the detected position, highly accurate position detection is realized even when the shape of the object changes over time.

In step S<b>105 , the positioning arm 103 is controlled by the position control section 114 based on the detected position thus obtained, and the welding section 102 is positioned with respect to the copper wire 301 . After that, the laser beam 105 is applied to the copper wire 301 under the control of the welding part 102 by the control part 104 to perform welding. In addition, the positioned part may be called a "welding part."

As described above, the cut shape of the copper wire 301 changes over time. Therefore, if the plurality of registered images 121 stored in the image storage unit 115 remain fixed, the accuracy of position detection deteriorates over time. Therefore, the image registration unit 116 automatically updates the registered image.

<Procedure for updating registered images>
FIG. 7 is a flowchart showing a processing procedure for updating a registered image by the image registration unit 116. As shown in FIG.
The registration image update process shown in the flowchart of FIG. 7 is executed, for example, when welding for one day is completed. It is assumed that the photographed images 310 for each welding time for one day of welding have already been stored in the image storage unit 115 .

When the registered image update process is started, in step S201, an average image (only the copper wire 301 portion) of the photographed images 310 for one day is created by alignment based on the detected position. Further, in step S202, position detection of the copper wire 301 is executed for each photographed image 310 for one day by pattern matching with the average image.

Thereafter, the processing (refining processing) of steps S201 and S202 is repeated a predetermined number of times (step S203; No). In the second and subsequent refining processes, alignment is performed in step S201 with reference to the detected position in the previous refining process. By performing the refining process, positional variations between the captured images 310 are corrected, blurring of the average image is reduced, and an average image with a clear pattern is obtained.

In other words, the refining process corresponds to the process of the average image updating section according to the present invention. That is, in the refinement process, the image registration unit 116 performs pattern matching between the average image and each of the plurality of photographed images 310 used to create the average image (step S201). Then, in refinement processing, the image registration unit 116 adjusts the position of each photographed image 310 based on the matching position, and creates a new average image from the plurality of photographed images 310 after position adjustment (step S202).

When the refining process is completed a predetermined number of times (step S203; Yes), in step S204, position detection in the procedure of steps S102 to S104 in FIG. It is executed for the captured image 310 . Then, the position detection result based on the registered image is compared with the position detection result based on the final average image as correct data, and the accuracy rate is obtained (step S205). That is, in step S205, pattern matching is performed between each of the plurality of photographed images 310 and a new registered image (for example, an average image) based on the plurality of photographed images 310 of the object (for example, the workpiece 300) obtained in the second time interval. A plurality of matching positions obtained in (1) are used as correct data.

Then, in step S205, for each photographed image 310, if the difference between the position detection result based on the registered image and the correct data is within the range of the copper wire 301, it is assumed that the position detection result based on the registered image is correct. , the accuracy rate for the photographed images 310 for one day is obtained.

If the obtained accuracy rate reaches the predetermined accuracy rate (step S205; Yes), updating of the registered image is unnecessary, and the process of updating the registered image ends. On the other hand, when the accuracy rate does not reach the predetermined accuracy rate (step S205; No), the registered image needs to be updated, and the average image finally obtained by the refining process is used as the new registered image in the image storage unit. 115 (step S206).

That is, the image registration unit 116 replaces the low-rate registered image with the new registered image when the matching rate between the correct data and the detection position based on the existing registered image is lower than the reference matching rate. As a result, the replacement of registered images automatically proceeds.

When storing a new registered image, the registered image with the lowest accuracy rate among the existing registered images is deleted from the image storage unit 115, and the process of updating the registered image ends. In other words, the process of step S206 corresponds to the process of the registered image updating unit according to the present invention, and the rate of matching between the pattern matching result for each of the plurality of photographed images 310 among the plurality of existing registered images 121 and the correct data is calculated. The low-rate enrollment image with the lowest is replaced with the new enrollment image.

FIG. 5 shows examples 122 and 123 of the accuracy rate for each registered image. In a first example 122 showing the accuracy rate for correct data on a certain day, the accuracy rate for the registered images a, b, e, and f is high. . On the other hand, in the second example 123 showing the accuracy rate for correct data on another day, the accuracy rate of registered images c, d, and f is high.

By registering registered images 121 with different patterns, even if a change such as fluctuation occurs in the cutting shape of the copper wire 301 , if the change is within the pattern in the plurality of registered images 121 , the plurality of registered images 121 can be used. Accurate position detection is possible by consensus.

However, if the cut shape of the copper wire 301 changes over time, the correct answer rate will decrease even by discussion. In that case, the registration image 121 is automatically updated according to the procedure shown in FIG. For example, in the case of the second example 123 of the accuracy rate, since the accuracy rate in the registered image a is "0.00%", the registered image a is deleted and the average image as a new registered image is stored. .

By automatically updating the registered image by the above procedure, the accuracy rate is improved, and the accuracy of position detection is maintained even when the cut shape of the copper wire 301 changes with time. However, when the wear of the blade 200 progresses and the blade 200 is replaced, the cut shape of the copper wire 301 undergoes a significant change that is different from aging and fluctuation.

In this case, as shown in FIG. 2, a user's reset instruction is given to the image registration unit 116, and all existing registered images 121 are deleted except, for example, the latest one. Then, until the number of registered images 121 reaches a predetermined number (six in the example of FIG. 5), the average image obtained by the refining process in steps S201 to S203 in FIG. added to.

That is, the image registration unit 116 functions as an example of the registration reset unit according to the present invention, and when receiving a predetermined instruction, the image registration unit 116 leaves all or part of the registered images 121 stored in the image storage unit 115. All others are discarded and a new registered image is stored. By resetting the registered image 121, the position detection by the control unit 104 can quickly cope with a large change in the cut shape of the copper wire 301 due to replacement of the blade 200 or the like.

Here, a welding device is given as an example of the usage of the position detection device, the positioning device, the position detection program, and the position detection method of the present invention. The method of use of the detection method is not limited to the above, and can be used in a wide range of applications such as machine tools and optical equipment.

Further, the position detection device of the present invention is not limited to a form in which a program is installed in a computer, and may be in a form in which the functional structure shown in FIG. 2 is realized by hardware.

The above-described embodiments should be considered illustrative and not restrictive in all respects. The scope of the present invention is indicated by the scope of the claims rather than the above-described embodiments, and is intended to include all modifications within the scope and meaning equivalent to the scope of the claims.

100: Welding device 101: Camera 102: Welding part 103: Positioning arm 104: Control part 105: Laser light 200: Blade 300: Work 301: Copper wire 111: Image acquisition part 112: Pattern matching part 113: Conference part 114: Position Control unit 115 : Image storage unit 116 : Image registration unit 302 : Cut edge 303 : Cut surface 305 : Crimped portion 310 : Photographed image 121 : Registered image 131 : Matching position 132 : Cluster 133 : Average position

Claims (11)

an image storage unit that stores a registered image for each first time interval based on the photographed image of the object in each of the plurality of first time intervals;
performing pattern matching between the registered image for each of the first time intervals and the photographed images of the object obtained in a second time interval different from the plurality of first time intervals, and performing pattern matching on the object for each of the registered images; a matching unit for obtaining a matching position of an object;
a position determination unit that determines the detection position of the object based on the matching position for each of the registered images;
A position detection device comprising: 2. The position detection according to claim 1, wherein the position determination unit determines, as the detection position, an average of the matching positions in a cluster to which the largest number of matching positions belong among the clusters into which the matching positions for each registered image are divided. Device. 3. The position detection device according to claim 1, wherein said registered image is an average image of a plurality of photographed images obtained within said first time interval. Pattern matching is performed between the average image and each of the plurality of captured images used to create the average image, position adjustment is performed for each captured image based on the matching position, and a plurality of captured images after position adjustment are performed. 4. The position detection device according to claim 3, further comprising an average image updating unit that creates a new average image from the average image. A plurality of matching positions obtained by pattern matching between each of the plurality of photographed images and a new registered image based on the plurality of photographed images of the object obtained in the second time interval are set as correct data, and the plurality of 2. A registered image updating unit that replaces a low-rate registered image with the lowest matching rate between a pattern matching result of each of said plurality of photographed images and said correct data among registered images of said new registered image with said new registered image. 5. The position detection device according to any one of 4 to 4. 6. The position according to claim 5, wherein the registered image update unit replaces the low-ratio registered image with the new registered image when a matching rate between the correct data and the detected position is lower than a reference matching rate. detection device. Claims comprising a registration reset section for discarding all or part of the registered images stored in the image storage section and storing a new registered image when a predetermined instruction is received. Item 7. The position detection device according to any one of Items 1 to 6. A position detection device according to any one of claims 1 to 7;
a positioning mechanism that positions an object to be positioned with reference to the detected position obtained by the position detecting device;
A positioning device comprising: A position detection device according to any one of claims 1 to 7;
a positioning mechanism that positions a welded portion based on the detected position obtained by the position detecting device;
a welding portion that welds the welding portion positioned by the positioning mechanism;
Welding equipment with A position detection program that causes a computer to function as each component in the position detection device according to any one of claims 1 to 7. A registered image for each first time interval based on a photographed image of an object in each of a plurality of first time intervals, and an image of the object obtained in a second time interval different from the plurality of first time intervals. a matching process of performing pattern matching with an image to obtain a matching position of the object for each registered image;
a position determination process of determining a detection position of the object based on the matching position for each of the registered images;
location detection method via

Images