Position detection of connection partners

Abstract

The invention relates to a method for determining the position of a connection partner for a connection process, comprising the following steps: providing (S10) the connection partner at the connection position; recording (S20) a real image of the positioned connection partner; generating (S30) a false color image based on the real image; and ascertaining (S40) at least one piece of information about the position of at least one of the connection partners on the basis of the false color image. The invention also relates to an apparatus and a computer program product for implementing the method.

Description

Technical Field

[0001] The present invention relates to the field of connection. In particular, the present invention relates to a method for determining the position of a connection partner for a connection process. Background Art

[0002] When welding component connections, it is important to know the exact position of the connection partners to be welded. Based on the position information, the welding path (especially the position of the welding path) can be adjusted precisely in order to create the best possible welded connection. Clamps are usually used to roughly position the connection partners. However, even with existing clamps, reliable, precisely repeatable positioning of the connection partners is often not possible. Therefore, in laser welding, computer programs with position detection algorithms are used for accurate position detection, which recognize the exact position or positioning of the connection partners relative to each other based on the camera image. However, the recognition can only be carried out accurately if the image area of ​​the camera image that is relevant for the position detection stands out with high contrast to the surrounding image area.

[0003] For example, in the so-called hairpin technology for producing stators for electric motors or generators, the ends of two adjacent hairpins are welded together in order to create an inserted coil. To this end, the ends are positioned approximately in the connection position with the aid of a positioning device. In order to better determine the exact position of the hairpin ends relative to each other, it is known to use hairpins with smoothed end faces. Due to the sufficient contrast, the position of the smoothed end faces of the hairpins can be reliably identified using known position detection algorithms, as shown in FIG. 2 a. However, in many cases, the creation of a smoothed connection surface is associated with additional effort and therefore increased costs.

[0004] The invention is based on the object of increasing the possibilities for automated position detection of connection partners using known position detection algorithms. Summary of the invention

[0005] The underlying object of the invention is achieved by the subject matter of the independent claims. Further possible embodiments of the invention are specified in the dependent claims, the description and the drawings. The features, advantages and possible embodiments set out in the description for one of the subject matter of the independent claims are to be regarded at least analogously as features, advantages and possible embodiments of the corresponding subject matter of the other independent claims and any possible combination of the subject matter of the independent claims, if combined with one or more of the dependent claims.

[0006] According to a first aspect of the present invention, a method for determining the position of a connection partner for a connection process is provided. The method can preferably precede the connection process for connecting the connection partner and be used to prepare the connection process. In the first step of the method, a connection partner is provided at a connection position. For the approximate orientation of the connection partner, the connection partner can be fixed at the connection position by means of a positioning device. The positioning device usually allows space for movement of the connection partner, which is why the connection partner can have a small offset (e.g., a few millimeters) or a small distance (e.g., a gap of a few millimeters or a few tenths of a millimeter) relative to each other. In a second method step, a real image of the connection partner positioned at the connection position is recorded. In particular, a camera device can be used for this purpose. The real image is a raster image that depicts the field of view of the camera device. In the subsequent third method step, a false color image is generated from the real image. In the context of the present disclosure, a false color image should be understood as a digital raster image (also referred to as a pixel image), the pixels of which can only assume one of at least two predetermined colors, in particular black or white and / or gray tones between them. When using a variety of gray tones as colors in a false color image, it is generally important to ensure that each grayscale has a clearly recognizable contrast relative to each other color used. When only two colors are used, the false color image can be a binary image. The false color image can visually highlight information, especially information about the position and / or geometry of the connection partner, which is related to connecting the connection partner. Then (in the fourth method step) based on the false color image, at least one information about the position of at least one of the connection partners is obtained. In particular, the position of at least one of the connection partners can be obtained relative to the position and / or orientation of the connection optical device, especially the laser welding head. The reference point of the position of the connection optical device can be the tool center point (TCP) of the connection optical device in particular. For example, in order to obtain position information, or in addition to obtaining position information, the corresponding shape and / or size of the connection partner can also be obtained.

[0007] According to a preferred variant, a deep convolutional neural network (also referred to as “CNN” for short), in particular a modified U-Net, can be used to generate a false color image from a real image.

[0008] In order to process the image data more reliably even under irregular contrast conditions, according to the invention, an intelligent filter is provided before the image evaluation, in particular the position detection, which generates a false color image. In this false color image, features relevant to the connection process are shown in one color (for example white), and features not relevant to the connection process (for example regions of the clamp or the connection partner that are not relevant to the connection process) are shown in another color (for example black). It should be understood that more than two feature classes can also be represented by means of the false color image, wherein preferably a separate color is used for the representation of each feature class, each of which has a sufficiently clear contrast relative to any other color used. Due to the optimal contrast between relevant features and irrelevant features, the proposed method allows a particularly robust detection of the position or positioning of the connection partner in order to achieve an optimized preparation for the connection process. In particular, a computer program with known components and a position detection algorithm can be used to easily and automatically identify information about the position of the connection partner based on the false color image and to derive certain connection settings for the subsequent connection process.

[0009] Preferably, in the fifth method step, the relative positions of the connection partners can be verified based on the determined position information. In this case, method step four and method step five can be run essentially in parallel.

[0010] Furthermore, a warning can be output if, during the verification of the position information and / or geometry-related information of at least one of the connection partners, an impermissible position and / or an impermissible geometry of at least one of the connection partners for the connection of the connection partners is determined. For example, a warning can be output if the shape or size of at least one of the connection partners exceeds a predefined limit for the shape or size of the connection partner in question. Furthermore, a warning can be output, for example, if the connection partners have an impermissible offset relative to one another or are positioned too far apart. In addition to or as an alternative to outputting a warning in the case of a critical geometry or position of the connection partners, it can also be provided that a subsequent connection process cannot be performed without additional manual authorization.

[0011] The verification step can also include determining process settings for connecting the connection partner. In particular, based on the position information, a movement path or a position of a predetermined movement path can be determined along which the laser beam moves over the surface of the connection partner for local melting of the connection partner.

[0012] In a further method step, the method can include connecting the connection partners using the determined process settings in a laser welding process. In this case, the claimed method can be implemented as part of a laser welding process.

[0013] In particular, if the claimed method is implemented as part of a laser welding process, a real image can be recorded from a direction perpendicular to the processing plane. In other words, the real image can represent a top view of the processing point of the connection partner. The processing plane is usually located in the region of the surface of the connection partner and, during the laser welding process, extends approximately transversely to the propagation direction of the processing laser beam in the region of the beam focus of the processing laser beam. During the laser welding process, a real image can be recorded by means of a camera device through the beam path of the processing laser beam. By orienting the camera device used concentrically with the beam path of the processing laser beam, the corresponding real image always shows the field of view of the processing optical device. In this way, the described method can also be used to easily determine the relative position of the connection partner relative to the position of the processing optical device or relative to the position of the tool center point of the connection device used.

[0014] The false color image generated by the real image can preferably have a resolution corresponding to the resolution of the real image. In this regard, one of at least two predetermined colors can be assigned to each pixel of the real image to generate a false color image. The color assignment can preferably be based on the assignment of each pixel to a predetermined feature class. This means that at least two feature classes can be predetermined for generating a false color image. For example, a feature class can be defined as the surface of at least one of the connection partners, and the second feature class can be defined as the background. Using trained artificial intelligence (KI) to assign pixels to one of the defined classes, the probability that each pixel belongs to one of the predetermined feature classes can be determined. If the probability is greater than a predetermined threshold, the corresponding pixel can be assigned to the feature class in question. As described above, KI can preferably be implemented as a deep convolutional neural network, preferably a U-Net / U-Net derivative, which has been optimized / trained for the situation in question and the running time. It should be understood that more than two feature classes can also be predetermined for generating a false color image. For example, different connection partners or additionally the components of the positioning device can be assigned to different feature classes, and therefore assigned to different colors of the false color image.

[0015] For example, the connection partner may be workpieces to be welded together. For example, the connection partner may be a metallic, rod-shaped conductor element - a so-called hairpin (essentially a wire bent in a U-shape) which is welded together to form an electrical connection. In particular, the hairpin may be made, for example, of copper or a copper-containing material, or of aluminum or an aluminum-containing material.

[0016] According to another aspect of the present invention, a laser welding device for connecting two connecting partners is provided. The laser welding device includes a positioning device for positioning the connecting partner in a connecting position. The positioning device can be, for example, a holder that fixes the connecting partners in a specific position (with a certain amount of clearance) relative to each other. The laser welding device also includes: a laser beam source for providing a processing laser beam; and a laser welding head having a laser welding optical device for orienting and focusing the processing laser beam in the direction of the connecting partner. The laser beam source can be, for example, a solid-state laser, in particular a fiber laser or a disk laser. The laser beam can be guided at least partially from the laser beam source to the laser welding head by means of an optical fiber. For example, scanning optics or also fixed optics can be used as laser welding optics. The laser welding device also includes: an image detection device for detecting a real image of the connecting partner in the connecting position. In particular, the image detection device can be a camera, such as a grayscale camera with a resolution in the megapixel range. The laser welding device further comprises an image processing device for generating a false color image based on the real image, and a measuring device for determining at least one information about the position of at least one of the connection partners based on the false color image.

[0017] The laser welding device may further include: a verification device for verifying the position and / or geometry of at least one of the connection partners based on at least one position information. The image processing device, the measuring device and the verification device may each be implemented as a computer program (product) on a common computer, for example, the common computer may also be used to control the camera device.

[0018] The verification device can also be designed to output a warning. For example, if an inadmissible position and / or an inadmissible geometry of at least one of the connection partners for the connection of the connection partners is determined during the verification of the position and / or geometry, a warning can be output. In addition to or as an alternative to the warning, the verification device can be designed to send a signal to a control device of the laser welding device, which prohibits the welding process due to the inadmissible position or geometry information of the connection partners before the welding process is manually allowed.

[0019] The verification device can also be designed to determine, based on the at least one piece of position information, a process setting for laser welding of the connection partner by means of the laser welding device. In particular, the verification device can send instructions for positioning and moving the laser processing beam to a control device of the laser welding device based on the detected relative position of the connection partner with respect to the laser welding head.

[0020] According to a further aspect of the invention, a computer program product is provided, which contains computer-readable instructions for executing a method according to one of the above-described variants on a connecting device, in particular on a laser welding device according to one of the above-described variants. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The following description of preferred exemplary embodiments serves to explain the invention in more detail with reference to the accompanying drawings.

[0022] In the attached picture:

[0023] Figure 1 The sequence of the method according to the invention is schematically shown;

[0024] FIG. 2 shows a real image of two hairpins in a welding position, wherein the ends of the hairpins to be welded have uniform end surfaces;

[0025] Figure 3a A real image of two hairpins in a welding position is shown, wherein the ends of the hairpins to be welded have uneven end surfaces;

[0026] Figure 3b Shown based on Figure 3a A false color image (here a binary image) of the real image shown; and

[0027] Figure 4 A laser welding device according to the present invention is schematically shown.

[0028] In the figures, identical elements or elements with identical functions are identified by identical reference symbols. DETAILED DESCRIPTION

[0029] Figure 1 A schematic diagram of a method according to the invention for determining a connection partner 20a, 20b for a connection process (see, for example, Figure 4 ) is located in the order of the method. In a first step, the method comprises providing S10 the connection partner 20a, 20b in the connection position. In a second step S20, a real image of the located connection partner is recorded (see also according to Figure 3a In the third step S30, based on the real image 22 (see also Figure 3b In a fourth step S40, at least one information about the position of at least one of the connection partners 20a, 20b is determined based on the false color image 24.

[0030] In an optional fifth step S50 , the position and / or geometry of at least one of the connection partners 20 a , 20 b may be verified based on the information obtained in step S40 .

[0031] In an optional sixth method step S60 , a warning may be output if an impermissible position and / or an impermissible geometry of at least one of the connection partners 20a , 20b for the connection of the connection partners 20a , 20b is detected during the verification S50 of the position and / or geometry.

[0032] Preferably, based on the verified position and / or geometry information of the connection partners 20a, 20b, settings for the connection process can be automatically determined, and the connection partners 20a, 20b can be welded together using the determined settings in a seventh step S70, in particular in a laser welding process. In this regard, determining the settings for the connection process can in particular include determining the position (or location) of a predetermined welding path and / or determining the coordinates of the starting point of the welding path.

[0033] FIG2 shows a top view of the connection surfaces of two hairpins 20 a, 20 b to be welded together. The “hairpin” technique is a winding technique for stators in electric motors and generators, in which a plurality of substantially U-shaped conductor elements - so-called “hairpins” - are welded together in pairs at their ends in order to produce a coil winding. The name “hairpin” comes from the U-shaped bent conductor element, which resembles the shape of a hairpin. To weld the two hairpins 20 a, 20 b, a focused treatment laser beam L (see FIG. Figure 4) moves along a treatment path (e.g. in a circular pendulum motion) over the connection surface at the end of the hairpins 20a, 20b. The energy introduced by means of the treatment laser beam L puts the material of the hairpins 20a, 20b in a molten state, so that a common molten bead is formed, which connects the two hairpins 20a, 20b to each other. Since more than 200 pairs of hairpins may need to be welded for an electric machine - depending on the design - it is very important to carry out this process in a fast and robust manner. In order to obtain a good welding result, it is important that the welding process and in particular the position of the welding path match the position of the ends of the hairpins to be connected. Therefore, for the precise position detection of the hairpins 20a, 20b, an image processing program is used, which determines the position of the hairpins 20a, 20b, in particular relative to the position or orientation of the welding optical device used, based on the recorded images. Based on the detected position of the connection partner 20a, 20b, the position and / or the starting point of the welding path are then adjusted for the welding process. For automated image processing, a high-contrast image of the hairpin ends is required, in which the contour of the hairpin ends associated with the welding stands out from the background (e.g. other hairpin areas, clamps, etc.). Therefore, it is known from the related art to cut the ends of the hairpins smoothly so that they have a surface that is as bare and smooth as possible. This surface reflects the light illuminated by the camera device and stands out clearly from the environment in the real image 22 of the camera device, so that the generated real image 22 can be used for automatic position recognition with the help of conventional position detection algorithms.

[0034] Since the production of hairpins 20a, 20b with exposed, uniform end faces requires complex additional processing steps (e.g., cutting off the hairpin ends using expensive cutting systems), the present invention proposes a solution by means of which uneven or contaminated connection partners (e.g., according to Figure 3a The hairpin ends) can also be reliably and accurately welded together using known position detection algorithms. Figure 3a , Figure 3b and Figure 4 The solution according to the invention is explained in more detail using the example of a hairpin. Figure 3a A real image 22 of the ends of two hairpins 20a, 20b positioned in a positioning device 12 for laser welding is shown. In this case, the ends of the hairpins 20a, 20b have a top-shaped structure, which is formed due to the previous processing steps of the hairpins 20a, 20b. Due to the surface quality of the hairpin ends, according to Figure 3a The real image 22 does not exhibit a sufficiently strong contrast between those image regions 222 showing the hairpin ends and the adjacent image regions 224, which is why the position of the hairpin ends relative to each other cannot be reliably determined using conventional position detection algorithms.

[0035] It is also known from the related art to record the real image 22 by means of a camera 18 of the laser welding system 10 (for example, a monochrome industrial camera with a resolution of 1.5 MPx), which is oriented coaxially with the process laser beam L (see Figure 4 ). In particular, in this respect, the optical axis of the camera 18 is coupled via a partially transparent mirror to the welding optics 162 in the laser welding head 16 of the laser welding device 10. In this way, the camera image always shows the processing area of ​​the processing laser beam L. In turn, the processing laser beam L is provided from the laser beam source 14 to the laser welding head 16 via an optical fiber cable and is focused on the ends of the hairpins 20a, 20b via the welding optics 162, which can be designed, for example, as a scanning optics with a movable scanning mirror or lens. The laser welding device 10 can be controlled by means of a central control device 19.

[0036] The real image 22 recorded by the camera device 18 (see Figure 3a ) is transmitted to a deep convolutional neural network of the image processing device 182, which processes the real image 22 and generates as output a false color image 24, here in the form of a binary image 24, with the same resolution as the real image 22. The preferred neural network for image conversion is a U-Net / U-Net derivative, which has been optimized for the case in question and the running time. In the generated binary image 22, the features related to the welding process are highlighted. This means that those pixels in the real image 22 that can be assigned as a feature class (with sufficient probability) to the surface of one of the hairpins 20a, 20b are painted white, and those pixels in the real image 22 that can be assigned to the "background" feature class are painted black. Therefore, the image area 242 that is identified as the surface of the hairpin 20a, 20b is shown in white and clearly stands out from the black background 244.

[0037] Now, according to the conventional process, the binary image 24 is transmitted to the measuring device 182 instead of the real image 22, which determines information about the position and / or geometry of the hairpins 20a, 20b. Then, with the help of the verification device 182, the obtained information can be verified for its suitability for the welding process, and the welding settings (especially the shape and / or position of the welding path to be traversed by the laser beam L) can be determined for the laser welding process based on the position information. For this purpose, the verified information about the position of the hairpin ends can be transmitted from the verification device to the central control device 19, which controls the laser welding device 10 according to the predefined welding settings. When verifying the position and / or geometry of the hairpin ends, it can be verified in particular whether the predefined position limits of the hairpin ends are maintained, whether the gap between the hairpins 20a, 20b is within a predefined tolerance range, or whether the cross section of the hairpins 20a, 20b exceeds or falls below a predefined size range. If it is determined in this regard that the predefined tolerance limits are not maintained, the verification device 182 can cooperate with the control device 19 to output a warning signal and / or stop the ongoing connection process.

[0038] according to Figure 4 The structure of the laser welding device 10 shown, the image processing device, the measuring device and the monitoring device can be implemented in a common computer unit 182. The computer unit 182 can be a tablet computer for image monitoring. The applicant sells an image monitoring system under the name "VisionLine" that can be used to implement the present invention.

[0039] In the system described herein, the neural network used by the image processing device 182 can be considered as an intelligent filter that highlights the features of the recorded real image 22 that are relevant to the welding process for automated image processing, thereby simplifying position detection. In this regard, the detection is not limited to one feature or feature class, but any number of feature classes that can be detected in the image and highlighted in a single false color image using two or more colors to reduce processing time. By determining the feature area (in this case, the surface of the hairpin end) with pixel accuracy, downstream algorithms can still perform reliability checks and compare the calculated / detected area with the expected area. This is difficult to achieve using classical intensity-based algorithms and can also have a positive impact on the robustness of the process.

[0040] According to the present invention, the position detection of the connection partners 20a, 20b allows reliable connection of the connection partners 20a, 20b. In addition, in some processes, the workload (cost) involved in pre-processing the connection partners 20a, 20b (e.g., accurate cutting of the hairpin ends) can be reduced.

Claims

1. A method for determining the position of a connection partner (20a, 20b) for a connection process, the method comprising the following steps: - providing (S10) the connection partner (20a, 20b) in a connection position; - recording (S20) a real image (22) of the positioned connection partners (20a, 20b); - generating (S30) a false color image (24) based on the real image (22); as well as - based on the false color image (24), determining (S40) at least one piece of information about the position of at least one of the connection partners (20a, 20b).

2. The method according to claim 1, in, The false color image (24) is generated from the real image (22) using a deep convolutional neural network, in particular a modified U-Net.

3. The method according to claim 1 or 2, further comprising: - Based on at least one piece of position information, verifying (S50) the relative position of the connection partners (20a, 20b) to each other.

4. The method according to claim 3, further comprising: - If, when verifying (S50) the position information or geometry-related information of at least one of the connection partners (20a, 20b), an inadmissible position and / or an inadmissible geometry of at least one of the connection partners (20a, 20b) for the connection of the connection partners (20a, 20b) is determined, a warning is output (S60).

5. The method according to claim 3 or 4, in, The step of verifying (S50) includes determining processing settings for connecting the connection partners (20a, 20b).

6. The method according to claim 5, further comprising: - Connecting the connection partners (20a, 20b) using the determined process settings in a laser welding process (S70).

7. The method according to any one of the preceding claims, in, The real image (22) is recorded from a direction perpendicular to the processing plane.

8. The method according to any one of the preceding claims, in, The false color image (24) has a resolution corresponding to the resolution of the real image (22).

9. The method according to any one of the preceding claims, in, One of at least two predetermined colors, in particular black or white and / or at least one gray tone therebetween, is assigned to each pixel of the real image (22) to generate the false color image (24).

10. The method according to claim 9, in, The color assignment is based on the allocation of individual pixels to predefined feature classes.

11. The method according to any one of the preceding claims, in, The connection partners (20a, 20b) are metallic, rod-shaped conductor elements, so-called hairpins, which are welded together to form an electrical connection.

12. A laser welding device (10) for connecting two connecting partners (20a, 20b), the laser welding device (10) comprising: - a positioning device for positioning the connection partner in the connection position; - a laser beam source (14) for providing a processing laser beam (L); a laser welding head (16) having laser welding optics (162) for orienting and focusing the processing laser beam (L) in the direction of the connection partner (20a, 20b); - an image detection device (18) for detecting a real image (22) of the connection partner (20a, 20b) in the connection position; - an image processing device (182) for generating a false color image (24) based on the real image (22); and A measuring device (182) for determining at least one piece of information about the position of at least one of the connection partners (20a, 20b) based on the false color image (24).

13. The laser welding device (10) according to claim 12, further comprising: - A verification device (182) for verifying the position and / or geometry of at least one of the connection partners (20a, 20b) based on the at least one position information, wherein the verification device (182) is also configured to output a warning if, when verifying the position and / or the geometry, an inadmissible position and / or an inadmissible geometry of at least one of the connection partners (20a, 20b) is determined for the connection of the connection partners (20a, 20b).

14. The laser welding device (10) according to claim 13, in, The verification device (182) is further configured to determine, based on the at least one piece of position information, process settings for laser welding the connection partners (20a, 20b) by means of the laser welding device (10).

15. A computer program product comprising computer readable instructions for executing the method according to any one of claims 1 to 11 on a connection device.

Images