Welding Detection Method and Related System in the Automobile Processing

Abstract

The present application provides a welding detection method and related system in the process of automobile processing. The method includes: extracting first features from a first welding area image to obtain a first feature set, matching the first feature set with a first reference feature set to obtain a first matching value; determining a first welding parameter and a first welding detection standard parameter corresponding to the first welding position; performing multi-scale decomposition on the first welding area image according to a first multi-scale decomposition algorithm and a first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; extracting features from the m high-frequency component parts to obtain k types of high-frequency feature parameters; matching the k types of high-frequency feature parameters with k types of standard high-frequency feature parameters to obtain a second matching value; determining a target matching value according to the first matching value and the second matching value. Using the present application can improve the welding detection efficiency in the process of automobile processing.

Description

Technical Field

[0001] The present application relates to the field of automotive technology, and in particular to a welding detection method and a related system during automobile processing. Background Art

[0002] In the automotive processing industry, the quality of welding directly affects the performance of the vehicle. In practical applications, after welding is completed, the welding inspection process is also very important. At present, the welding inspection efficiency is low. Therefore, how to improve the welding inspection efficiency in the automotive processing process needs to be solved urgently. Summary of the invention

[0003] The embodiments of the present application provide a welding detection method and a related system during automobile processing, which can improve the welding detection efficiency during automobile processing.

[0004] In a first aspect, an embodiment of the present application provides a welding detection method in an automobile processing process, which is applied to an industrial robot, wherein the industrial robot includes a visual system and a welding system, and the method includes:

[0005] Acquire a first image for a first welding position by the visual system, where the first welding position is a welded position between a first automobile component and a second automobile component;

[0006] Performing region segmentation on the first image to obtain a first welding region image, performing first feature extraction based on the first welding region image to obtain a first feature set, acquiring a first reference feature set corresponding to the first welding position, and matching the first feature set with the first reference feature set to obtain a first matching value;

[0007] When the first matching value is within a first preset range, the first welding parameter and the first welding detection standard parameter corresponding to the first welding position are determined by the welding system, the first welding detection standard parameter includes k standard high-frequency characteristic parameters, each standard high-frequency characteristic parameter corresponds to at least one high-frequency characteristic type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold;

[0008] determining a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter;

[0009] The first welding area image is multi-scale decomposed according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer;

[0010] Feature extraction is performed on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameters to obtain k types of high-frequency feature parameters;

[0011] The k types of high-frequency feature parameters are matched with the k types of standard high-frequency feature parameters to obtain a second matching value;

[0012] A target matching value is determined according to the first matching value and the second matching value;

[0013] When the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding detection at the first welding position is qualified.

[0014] In a second aspect, an embodiment of the present application provides a welding detection system during automobile processing, which is applied to an industrial robot. The industrial robot includes a vision system and a welding system. The welding detection system during automobile processing includes:

[0015] An acquisition unit, configured to acquire a first image of a first welding position through the vision system, where the first welding position is a welded position between a first automobile part and a second automobile part;

[0016] A segmentation unit, configured to perform region segmentation on the first image to obtain a first welding region image, perform first feature extraction according to the first welding region image to obtain a first feature set, acquire a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value;

[0017] A determination unit, configured to, when the first matching value is within a first preset range, determine, through the welding system, a first welding parameter and a first welding detection standard parameter corresponding to the first welding position. The first welding detection standard parameter includes k types of standard high-frequency feature parameters, and each type of standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper limit threshold and a first lower limit threshold; determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter;

[0018] A decomposition unit, configured to perform multi-scale decomposition on the first welding region image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer;

[0019] An extraction unit, configured to perform feature extraction on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter, so as to obtain k types of high-frequency feature parameters;

[0020] A matching unit, configured to match the k types of high-frequency feature parameters with the k types of standard high-frequency feature parameters to obtain a second matching value;

[0021] The determination unit is configured to determine a target matching value according to the first matching value and the second matching value; when the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding detection at the first welding position is qualified.

[0022] In a third aspect, an embodiment of the present application provides an industrial robot, including a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the processor, and the programs include instructions for executing the steps in the first aspect of the embodiments of the present application.

[0023] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program for electronic data exchange, and the computer program enables a computer to execute some or all of the steps described in the first aspect of the embodiments of the present application.

[0024] In a fifth aspect, an embodiment of the present application provides a computer program product, wherein the computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute some or all of the steps described in the first aspect of the embodiments of the present application. The computer program product may be a software installation package.

[0025] Implementing the embodiments of the present application has the following beneficial effects:

[0026] It can be seen that the welding detection method and related system in the automobile processing process described in the embodiments of the present application are applied to an industrial robot, which includes a vision system and a welding system. The vision system acquires a first image of a first welding position, where the first welding position is the welded position between a first automobile component and a second automobile component; the first image is subjected to region segmentation to obtain a first welding region image, and first feature extraction is performed on the first welding region image to obtain a first feature set. A first reference feature set corresponding to the first welding position is acquired, and the first feature set is matched with the first reference feature set to obtain a first matching value; when the first matching value is within a first preset range, the welding system determines a first welding parameter and a first welding detection standard parameter corresponding to the first welding position. The first welding detection standard parameter includes k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter are determined; the first welding region image is subjected to multi-scale decomposition according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; feature extraction is performed on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter to obtain k high-frequency feature parameters; the k high-frequency feature parameters are matched with the k standard high-frequency feature parameters to obtain a second matching value; a target matching value is determined according to the first matching value and the second matching value;When the target matching value is greater than or equal to the first upper threshold, it is determined that the welding detection at the first welding position is qualified. First, the characteristics of the qualified solder joints corresponding to the welding position can be obtained, and the first feature set is matched with the first reference feature set to obtain the first matching value, realizing welding detection at the appearance level initially. Second, when the first matching value is within the first preset range, it indicates that the welding appearance is average. The first welding parameters and the first welding detection standard parameters corresponding to the first welding position can be determined through the welding system. In this way, the welding parameters corresponding to the welding position and the corresponding welding detection standard parameters can be obtained. Third, not only can the multi-scale decomposition algorithm corresponding to the welding process be adopted, which can ensure the in-depth extraction of the welding features corresponding to the welding process. Additionally, the corresponding algorithm control is also based on the welding detection standard, which helps to in-depth extract the welding features corresponding to the welding process, thus helping to improve the welding detection efficiency in the automotive processing process. Fourth, the low-frequency component part reflects the main body of the image, and m high-frequency component parts reflect the details of the image. The low-frequency component part and the m high-frequency component parts reflect the influence degree of the main body of the image on the image details. The first welding detection standard parameter reflects the welding detection standard corresponding to the welding process, that is, the corresponding high-frequency feature parameters can be extracted based on the influence degree of the main body of the image on the image details (the characteristics of the image itself) and the welding detection standard corresponding to the welding process, so that the welding features corresponding to the welding process can be in-depth extracted. Fifth, the welding quality is deeply evaluated based on the welding features at the appearance level and the deep welding details, that is, the target matching value is determined according to the first matching value and the second matching value. When the target matching value is greater than or equal to the first upper threshold, it indicates that the welding is good, that is, it is determined that the welding detection at the first welding position is qualified, thus helping to improve the welding detection efficiency in the automotive processing process. Brief Description of the Drawings

[0027] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0028] Figure 1 It is a flowchart showing the welding detection method in the automotive processing process provided by the embodiment of the present application;

[0029] Figure 2 It is a structural schematic diagram of an industrial robot provided by the embodiment of the present application;

[0030] Figure 3 It is a block diagram showing the functional unit composition of the welding detection system in the automotive processing process provided by the embodiment of the present application. Detailed implementation manners

[0031] In the description and claims of this application and the above-mentioned drawings, terms such as "first", "second", etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but may also include steps or units not listed in a possible example, or may also include other steps or units inherent to these processes, methods, products or devices in a possible example.

[0032] Referring to "embodiments" herein means that the specific features, structures or characteristics described in connection with the embodiments can be included in at least one embodiment of this application. The appearance of this phrase in various positions in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0033] In order to enable those skilled in the art of this technology to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope protected by this application.

[0034] In the embodiments of this application, the industrial robot may include various industrial robots with visual detection functions in the automobile processing process. For example, automobile processing robots, intelligent machine tools, humanoid robots, etc. are not limited herein.

[0035] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a welding detection method in the automobile processing process provided by the embodiments of this application, applied to an industrial robot. The industrial robot includes a vision system and a welding system. The welding detection method in the automobile processing process of this application includes:

[0036] 101. Obtain a first image of a first welding position through the vision system, where the first welding position is the welded position between a first automobile part and a second automobile part.

[0037] Among them, the first automotive component and the second automotive component can be automotive components of the same or different types. Both the first automotive component and the second automotive component can be understood as automotive components that need to be welded, and can include at least one of the following: frame body, battery busbar, screw, car door, chassis, axle, nut, etc., which are not limited herein.

[0038] Among them, the vision system can include at least one vision sensor, and the vision sensor can include at least one of the following: camera, ultrasonic sensor, radar sensor, infrared sensor, etc., which are not limited herein.

[0039] Among them, the welding system can include a welding torch, and the welding torch can record the welding parameters of each welding position during the welding process and implement the welding function.

[0040] In a specific implementation, a first image of a first welding position can be obtained through the vision system, and the first welding position is any welding position among the welded positions between the first automotive component and the second automotive component.

[0041] 102. Perform region segmentation on the first image to obtain a first welding region image, perform first feature extraction based on the first welding region image to obtain a first feature set, obtain a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value.

[0042] Among them, the first feature set can include one or more features, and the feature can include at least one of the following: solder joint color, average solder joint width, maximum solder joint width, minimum solder joint width, average solder joint thickness, maximum solder joint thickness, minimum solder joint thickness, etc., which are not limited herein.

[0043] Among them, the first reference feature set can include one or more features, and the feature can include at least one of the following: solder joint color, average solder joint width, maximum solder joint width, minimum solder joint width, average solder joint thickness, maximum solder joint thickness, minimum solder joint thickness, etc., which are not limited herein.

[0044] In a specific implementation, the first image can be subjected to region segmentation to obtain a first welding region image. The first welding region image can be understood as an image corresponding only to the welding region. In a specific implementation, the first welding region image can be subjected to image recognition to implement first feature extraction and obtain a first feature set.

[0045] In specific implementation, the welding processes corresponding to different welding positions are different, and for different welding processes, the characteristics of qualified welds are also different. The mapping relationship between the preset welding positions and the reference feature sets can be stored in advance. Furthermore, based on this mapping relationship, the first reference feature set corresponding to the first welding position can be determined. In this way, the characteristics of qualified solder joints corresponding to the welding positions can be obtained. Furthermore, the first feature set can be matched with the first reference feature set to obtain the first matching value. In this way, welding detection can be initially realized from the appearance level, and the welding detection efficiency in the automotive processing process can be improved.

[0046] 103. When the first matching value is within the first preset range, the welding system determines the first welding parameters and the first welding detection standard parameters corresponding to the first welding position. The first welding detection standard parameters include k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper limit threshold and a first lower limit threshold.

[0047] Among them, the first preset range can be set in advance or be the system default. The first preset range includes a first upper limit threshold and a first lower limit threshold, and the first lower limit threshold is less than the first upper limit threshold.

[0048] Among them, the welding system includes: a wire feeder, a welding power source, and a welding torch.

[0049] Among them, the first welding parameters corresponding to the first welding position may include at least one of the following: welding position coordinates, welding temperature, welding trajectory, welding rate of the welding torch, wire feeding rate of the wire feeder, power source parameters of the welding power source, welding action parameters, etc., which are not limited here.

[0050] Among them, the power source parameters of the welding power source may include at least one of the following: working current of the welding power source, working voltage of the welding power source, working power of the welding power source, working mode of the welding power source, etc., which are not limited here.

[0051] Among them, the welding action parameters may include at least one of the following: welding action amplitude, welding process, welding action trajectory, welding angle, etc., which are not limited here.

[0052] Among them, the standard high-frequency feature parameters can be understood as the feature parameters of the high-frequency component part, and the standard high-frequency feature parameters can include at least one of the following: feature points, eigenvalues, feature patterns, feature vectors, total number of feature points, feature point distribution density (the ratio between the total number of feature points and the area of the region where the feature is located), etc., which are not limited herein. In specific implementation, the first welding detection standard parameters can include k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer.

[0053] In specific implementation, different welding positions can correspond to different welding detection standard parameters, that is, the mapping relationship between the preset welding positions and the welding detection standard parameters can be stored in advance. Furthermore, based on this mapping relationship, the first welding detection standard parameters corresponding to the first welding position can be determined. In this way, the welding detection standard parameters corresponding to the welding process can be obtained.

[0054] Alternatively, different welding positions can also correspond to different sets of welding detection standard parameters, that is, the mapping relationship between the preset welding positions and the sets of welding detection standard parameters can be stored in advance. Furthermore, based on this mapping relationship, the first set of welding detection standard parameters corresponding to the first welding position can be determined. The first set of welding detection standard parameters can include multiple welding detection standard parameters, and each welding detection standard parameter can correspond to a welding parameter. Furthermore, the first welding detection standard parameter corresponding to the first welding parameter can be selected from the first set of welding detection standard parameters. In this way, the welding detection standard parameters corresponding to the welding position and the welding process can be obtained.

[0055] Among them, the first welding detection standard parameters can be understood as the welding detection standard parameters in the case of qualification corresponding to the first welding position. To a certain extent, the first welding detection standard parameters can guide which type of high-frequency features need to be extracted, or which layer of high-frequency features, or perform corresponding processing on one or more high-frequency features in one or more high-frequency component parts, such as taking the average value, quantity statistics, summation, etc.

[0056] In specific implementation, when the first matching value is within the first preset range, that is, the first matching value is greater than or equal to the first lower threshold and less than the first upper threshold, it indicates that the welding appearance is average. Then, the first welding parameter and the first welding detection standard parameters corresponding to the first welding position can be determined through the welding system. In this way, the welding parameters corresponding to the welding position and the corresponding welding detection standard parameters can be obtained. In this way, the welding detection efficiency in the automobile processing process can be improved.

[0057] 104. Determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter.

[0058] In specific implementation, the multi-scale decomposition algorithm may include at least one of the following: wavelet transform algorithm, contourlet transform algorithm, non-subsampled contourlet transform algorithm, Laplacian pyramid transform algorithm, etc., which is not limited here. Specifically, the mapping relationship between the preset welding parameters and the multi-scale decomposition algorithm can be stored in advance. Then, based on this mapping relationship, the first multi-scale decomposition algorithm corresponding to the first welding parameter is determined. That is, different welding parameters reflect different welding processes, and different welding processes require different welding characteristics to be detected. In this way, the multi-scale decomposition algorithm corresponding to the welding process can be adopted, so that the welding characteristics corresponding to the welding process can be deeply extracted.

[0059] In specific implementation, various algorithm control parameters of the first multi-scale decomposition algorithm can also be stored in advance. Each algorithm control parameter corresponds to a welding detection standard parameter. Then, the first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter can be determined. Among them, the first algorithm control parameter is used to control the algorithm effect of the first multi-scale decomposition algorithm. The algorithm effect may include at least one of the following: algorithm speed, number of layers of the high-frequency component part, multi-scale decomposition degree, etc., which is not limited here. That is, corresponding algorithm control based on the welding detection standard helps to deeply extract the welding characteristics corresponding to the welding process, and thus helps to improve the welding detection efficiency in the automobile processing process.

[0060] 105. Perform multi-scale decomposition on the first welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer.

[0061] In specific implementation, the first welding area image can be multi-scale decomposed according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer. Not only can the multi-scale decomposition algorithm corresponding to the welding process be adopted, so that the welding characteristics corresponding to the welding process can be deeply extracted. In addition, corresponding algorithm control is also performed based on the welding detection standard, which helps to deeply extract the welding characteristics corresponding to the welding process, and thus helps to improve the welding detection efficiency in the automobile processing process.

[0062] 106. Extract features from the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter to obtain k kinds of high-frequency feature parameters.

[0063] Among them, the low-frequency component part reflects the main body of the image, and the m high-frequency component parts reflect the details of the image. The low-frequency component part and the m high-frequency component parts reflect the influence degree of the main body of the image on the details of the image. In specific implementation, k kinds of high-frequency feature parameters are obtained by performing feature extraction on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameters. Since the first welding detection standard parameters reflect the welding detection standard corresponding to the welding process, that is, the corresponding high-frequency feature parameters can be extracted based on the influence degree of the main body of the image on the details of the image (the characteristics of the image itself) and the welding detection standard corresponding to the welding process. Therefore, the welding features corresponding to the welding process can be deeply extracted, which helps to improve the welding detection efficiency in the automotive processing process.

[0064] Optionally, step 106 above, performing feature extraction on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameters to obtain k kinds of high-frequency feature parameters may include the following steps:

[0065] Determine the first information entropy of the low-frequency component part;

[0066] Determine the second information entropy of the first welding area image;

[0067] Determine the ratio between the first information entropy and the second information entropy to obtain a first ratio;

[0068] When the first ratio is within a second preset range, perform feature extraction on the m high-frequency component parts according to the first welding detection standard parameters to obtain the k kinds of high-frequency feature parameters; the second preset range includes a second upper limit threshold and a second lower limit threshold.

[0069] Among them, the second preset range can be set in advance or be the system default. The second preset range includes a second upper limit threshold and a second lower limit threshold. The second lower limit threshold is less than the second upper limit threshold.

[0070] In a specific implementation, the first information entropy of the low-frequency component part can be determined, the second information entropy of the first welding area image can also be determined, and the ratio between the first information entropy and the second information entropy can be determined to obtain a first ratio. The first ratio = the first information entropy / the second information entropy. The first ratio reflects the influence degree of the image main body on the image details. When the first ratio is within a second preset range, that is, the first ratio is greater than or equal to a first lower limit threshold and less than or equal to a first upper limit threshold, it indicates that the relationship between the image main body and the image details meets the requirements. Since the first welding detection standard parameters can, to a certain extent, guide which type of high-frequency features need to be extracted, or which layer of high-frequency features, or perform corresponding processing on one or more high-frequency features in one or more high-frequency component parts, therefore, the k high-frequency feature parameters can be obtained by extracting features from the m high-frequency component parts according to the first welding detection standard parameters. In this way, the high-frequency feature parameters corresponding to the first welding detection standard parameters can be extracted, and further, the subsequent matching accuracy can be guaranteed, thereby helping to improve the welding detection efficiency in the automobile processing process.

[0071] Optionally, the following steps may further be included:

[0072] When the first ratio is greater than the second upper limit threshold, determine the third relative deviation degree between the first ratio and the second upper limit threshold, determine the first high-frequency image enhancement algorithm corresponding to the third relative deviation degree, and determine the first mean value between the first matching value and the second matching value; determine the first control parameter corresponding to the first high-frequency image enhancement algorithm corresponding to the first mean value;

[0073] Perform image enhancement on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter;

[0074] Extract features from the m high-frequency component parts after image enhancement according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0075] In a specific implementation, when the first ratio is greater than the second upper threshold, it indicates that the main body of the image obscures the image details. Furthermore, the third relative deviation between the first ratio and the second upper threshold can be determined, that is, the third relative deviation = (the first ratio - the second upper threshold) / (the first ratio + the second upper threshold). That is, the third relative deviation reflects the degree to which the main body of the image obscures the image details. The mapping relationship between the preset relative deviation and the high-frequency image enhancement algorithm can be pre-stored. The high-frequency image enhancement algorithm is used to implement high-frequency image enhancement. Thus, the significance of the image details can be enhanced, the integrity of the welding feature extraction and the quality of the welding features can be ensured. Furthermore, based on this mapping relationship, the first high-frequency image enhancement algorithm corresponding to the third relative deviation can be determined. In addition, the first matching value and the second matching value can be subjected to a mean operation, that is, the first mean = (the first matching value + the second matching value) / 2. The first mean represents the welding quality. Furthermore, the mapping relationship between the preset mean and the control parameters corresponding to the first high-frequency image enhancement algorithm can be pre-stored. Furthermore, based on this mapping relationship, the first control parameter corresponding to the first mean can be determined. The first control parameter is used to control the algorithm effect of the first high-frequency image enhancement algorithm. The algorithm effect can include at least one of the following: the degree of high-frequency image enhancement, the speed of high-frequency image enhancement, the area of high-frequency image enhancement, which high-frequency component part needs high-frequency image enhancement, which type of feature needs high-frequency image enhancement, etc., which is not limited here. In this way, the algorithm effect of the first high-frequency image enhancement algorithm can be dynamically adjusted based on the welding quality, so that the depth of the enhancement of the significance of the image details conforms to the actual situation.

[0076] Next, image enhancement is performed on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter. Thus, the significance of the image details can be enhanced, the integrity of the welding feature extraction and the quality of the welding features can be ensured. Since the first welding detection standard parameter can, to a certain extent, guide which type of high-frequency feature needs to be extracted, or which layer of high-frequency feature, or corresponding processing is performed on one or more high-frequency features in one or more layers of high-frequency component parts. Therefore, the k high-frequency feature parameters can be obtained by extracting features from the high-frequency component parts after image enhancement according to the first welding detection standard parameter. In this way, the high-frequency feature parameters corresponding to the first welding detection standard parameter can be extracted. Furthermore, the subsequent matching accuracy can be ensured, which helps to improve the welding detection efficiency in the automotive processing process.

[0077] Optionally, the following steps may further be included:

[0078] When the first ratio is less than the second lower threshold, determine the fourth relative deviation between the second upper threshold and the first ratio, determine the first low-frequency image enhancement algorithm corresponding to the fourth relative deviation, determine the first difference between the first matching value and the first lower threshold, and determine the second control parameter of the first low-frequency image enhancement algorithm corresponding to the first difference;

[0079] Perform image enhancement on the low-frequency component part according to the first low-frequency image enhancement algorithm and the second control parameter;

[0080] Reconstruct the second welding area image according to the image-enhanced low-frequency component part and the m high-frequency component parts;

[0081] Perform multi-scale decomposition on the second welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain m target high-frequency component parts;

[0082] Extract features from the m target high-frequency component parts according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0083] In specific implementation, when the first ratio is less than the second lower threshold, it indicates that the image main body is insufficient but the details are too prominent. Then, the fourth relative deviation between the second upper threshold and the first ratio can be determined. The fourth relative deviation = (second upper threshold - first ratio) / (second upper threshold + first ratio). The fourth relative deviation reflects the degree of over-prominence of the image details. The mapping relationship between the preset relative deviation and the low-frequency image enhancement algorithm can be pre-stored. The low-frequency image enhancement algorithm is used to implement low-frequency image enhancement, thereby enhancing the saturation of the image main body to suppress the prominence of the image details and ensure the accuracy of welding feature extraction. Furthermore, based on this mapping relationship, the first low-frequency image enhancement algorithm corresponding to the fourth relative deviation can be determined. Then, the first difference between the first matching value and the first lower threshold can be determined. The first difference = first matching value - first lower threshold. The appearance level is related to the depth of the image main body. Especially in terms of appearance, in general welding cases. Furthermore, based on the mapping relationship between the difference and the control parameter of the first low-frequency image enhancement algorithm, based on this mapping relationship, the second control parameter of the first low-frequency image enhancement algorithm corresponding to the first difference can be determined. The second control parameter is used to control the algorithm effect of the first low-frequency image enhancement algorithm. The algorithm effect can include at least one of the following: low-frequency image enhancement degree, low-frequency image enhancement speed, low-frequency image enhancement area, etc., which are not limited here. In this way, the algorithm effect of the first low-frequency image enhancement algorithm can be dynamically adjusted based on the quality of the appearance welding, so that the depth of the saturation increase of the image main body conforms to the actual situation.

[0084] Furthermore, the low-frequency component part can be enhanced according to the first low-frequency image enhancement algorithm and the second control parameter, so that the saturation improvement depth of the main body of the image conforms to the actual situation. Then, the second welding area image is reconstructed based on the enhanced low-frequency component part and the m high-frequency component parts, and the second welding area image is decomposed into m target high-frequency component parts according to the first multi-scale decomposition algorithm and the first algorithm control parameter. Then, k kinds of high-frequency feature parameters are obtained by extracting features from the m target high-frequency component parts according to the first welding detection standard parameter. Since the image detail saliency is too significant, after the image details are refilled, it is necessary to reconstruct first and then extract the high-frequency component parts. Therefore, the accuracy of welding feature extraction can be guaranteed. In addition, since the first welding detection standard parameter can guide to some extent which type of high-frequency feature needs to be extracted, or which layer of high-frequency feature, or perform corresponding processing on one or more high-frequency features in one or more layers of high-frequency component parts, k kinds of high-frequency feature parameters can be obtained by extracting features from the m high-frequency component parts according to the first welding detection standard parameter. In this way, high-frequency feature parameters corresponding to the first welding detection standard parameter can be extracted, and further, the subsequent matching accuracy can be guaranteed, thus helping to improve the welding detection efficiency in the automobile processing process.

[0085] 107. Match the k kinds of high-frequency feature parameters with the k kinds of standard high-frequency feature parameters to obtain a second matching value.

[0086] In specific implementation, each high-frequency feature parameter among the k kinds of high-frequency feature parameters can be matched with the corresponding standard high-frequency feature parameter among the k kinds of standard high-frequency feature parameters to obtain k matching values. The k matching values can be weighted to obtain the second matching value, that is, each standard high-frequency feature parameter among the k kinds of standard high-frequency feature parameters corresponds to a weight value. In this way, k weight values can be obtained, and the sum of the k weight values is 1. The k weight values can be preset or system default. Since the welding features corresponding to the welding process are deeply extracted and the corresponding welding features are matched with the standard welding features, the high-frequency feature verification can be realized through the image detail depth, and the welding detection result can be obtained, which helps to improve the welding detection efficiency in the automobile processing process.

[0087] 108. Determine the target matching value according to the first matching value and the second matching value.

[0088] In the embodiment of the present application, the welding quality can be evaluated based on the welding features at the appearance level and the deep welding detail depth, that is, determining the target matching value according to the first matching value and the second matching value, which helps to improve the welding detection efficiency in the automobile processing process.

[0089] Optionally, the first welding detection standard parameter further includes a standard matching value; step 108 above, determining the target matching value according to the first matching value and the second matching value may include the following steps:

[0090] Determine a first relative deviation degree between the first upper limit threshold and the first matching value;

[0091] When the first relative deviation degree is less than or equal to a preset relative deviation degree, determine a first deviation degree between the second matching value and the standard matching value;

[0092] Determine a first adjustment parameter corresponding to the first deviation degree;

[0093] Determine the target matching value according to the first adjustment parameter and the first matching value;

[0094] When the first relative deviation degree is greater than the preset relative deviation degree, determine a second relative deviation degree between the first matching value and the first lower limit threshold;

[0095] Determine a target weight pair corresponding to the second relative deviation degree, the target weight pair includes a first weight and a second weight, and the sum of the first weight and the second weight is 1;

[0096] Perform a weighted operation according to the first matching value, the second matching value, and the target weight pair to obtain the target matching value.

[0097] Among them, the first welding detection standard parameter may further include a standard matching value, and the standard matching value can be understood as a matching threshold. When the second matching value is greater than the standard matching value, it means that from the perspective of detailed features, the welding is good. On the contrary, when the second matching value is less than or equal to the standard matching value, it means that from the perspective of detailed features, the welding is average.

[0098] Among them, the preset relative deviation degree can be set in advance or default by the system.

[0099] In specific implementation, the first relative deviation between the first upper limit threshold and the first matching value can be determined. The first relative deviation = (the first upper limit threshold - the first matching value) / (the first upper limit threshold + the first matching value). The first relative deviation includes not only the relative deviation range but also the relative deviation direction. When the first relative deviation is less than or equal to the preset relative deviation, it indicates that the first matching value is very close to the first upper limit threshold. From the appearance, the welding is close to being good but has slight defects. Then, the welding verification can be further carried out by deeply combining the welding details, that is, the first deviation between the second matching value and the standard matching value can be determined. The first deviation = (the second matching value - the standard matching value) / the standard matching value. The first deviation includes not only the deviation magnitude but also the deviation direction, and the first deviation reflects the quality of the welding details.

[0100] Next, the mapping relationship between the preset deviation and the adjustment parameter can be pre-stored. Among them, the value range of the adjustment parameter can be preset or defaulted by the system. For example, the value range of the adjustment parameter is -0.2 to 0.2. Furthermore, the first adjustment parameter corresponding to the first deviation can be determined based on this mapping relationship, and then the target matching value can be determined according to the first adjustment parameter and the first matching value, that is, the target matching value = (1 + the first adjustment parameter) * the first matching value. Since the welding is close to being good from the appearance and the matching result (the first matching value) at the appearance level is considered completely credible, but there are slight defects, it indicates that the recognition of the welding details can deeply influence the welding verification result. Therefore, the credible matching result (the first matching value) can be dynamically adjusted by using the deviation degree between the actual welding details and the standard matching value, so that the final matching value deeply conforms to the actual situation, which helps to improve the welding detection efficiency in the automotive processing process.

[0101] Correspondingly, when the first relative deviation is greater than the preset relative deviation, from the appearance, the welding deviates from being good, that is, the welding is average. It is considered that the matching result (the first matching value) at the appearance level is not completely credible, but the credibility (weight) is between 0 and 1. Then, the second relative deviation between the first matching value and the first lower limit threshold is determined. The second relative deviation includes not only the relative deviation range but also the relative deviation direction. The second relative deviation = (the first matching value - the first lower limit threshold) / (the first matching value + the first lower limit threshold). The mapping relationship between the preset relative deviation and the weight pair can also be pre-stored. The weight pair includes two weights, one weight corresponds to the first matching value, and the other weight corresponds to the second matching value. Furthermore, the target weight pair corresponding to the second relative deviation can be determined based on this mapping relationship. The target weight pair includes the first weight and the second weight, and the sum of the first weight and the second weight is 1. Among them, the larger the second relative deviation, the larger the first weight, and the smaller the second relative deviation, the smaller the first weight.

[0102] Next, perform a weighted operation based on the first matching value, the second matching value, and the target weight pair to obtain the target matching value, that is, target matching value = first weight * first matching value + second weight * second matching value. In this way, in terms of appearance, when the welding deviation is good, that is, in the case of general welding, the corresponding weight pair is dynamically determined based on the deviation degree from the first lower threshold. That is, in the case of general welding, the detailed features are used to compensate for the welding deficiency in appearance. Thus, the final welding detection result conforms to the actual situation in depth, which helps to improve the welding detection efficiency in the automotive processing process.

[0103] 109. When the target matching value is greater than or equal to the first upper threshold, it is determined that the welding detection of the first welding position is qualified.

[0104] In specific implementation, when the target matching value is greater than or equal to the first upper threshold, it indicates that the welding is good, that is, it is determined that the welding detection of the first welding position is qualified. Thus, it helps to improve the welding detection efficiency in the automotive processing process.

[0105] Optionally, the following steps may also be included:

[0106] When the first matching value is greater than the first upper threshold, it is determined that the welding detection of the first welding position is qualified;

[0107] Or,

[0108] When the first matching value is less than the first lower threshold, it is determined that the welding detection of the first welding position is unqualified;

[0109] Or,

[0110] When the target matching value is less than the first upper threshold, it is determined that the welding detection of the first welding position is unqualified.

[0111] In specific implementation, when the first matching value is greater than the first upper threshold, it indicates that the welding is good, and it is determined that the welding detection of the first welding position is qualified.

[0112] In specific implementation, when the first matching value is less than the first lower threshold, it indicates that there are welding defects, that is, it can be determined that the welding detection of the first welding position is unqualified.

[0113] In specific implementation, when the target matching value is less than the first upper threshold, it indicates that there are welding defects, that is, it can be determined that the welding detection of the first welding position is unqualified.

[0114] It can be seen that the welding detection method in the automobile processing process described in the embodiments of the present application is applied to an industrial robot, which includes a vision system and a welding system. The vision system is used to obtain a first image of a first welding position, where the first welding position is the welded position between a first automobile component and a second automobile component; the first image is subjected to region segmentation to obtain a first welding region image, and first feature extraction is performed according to the first welding region image to obtain a first feature set. A first reference feature set corresponding to the first welding position is obtained, and the first feature set is matched with the first reference feature set to obtain a first matching value; when the first matching value is within a first preset range, the welding system is used to determine a first welding parameter and a first welding detection standard parameter corresponding to the first welding position. The first welding detection standard parameter includes k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter are determined; the first welding region image is subjected to multi-scale decomposition according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; feature extraction is performed on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter to obtain k high-frequency feature parameters; the k high-frequency feature parameters are matched with the k standard high-frequency feature parameters to obtain a second matching value; a target matching value is determined according to the first matching value and the second matching value;When the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding detection at the first welding position is qualified. Firstly, the characteristics of the qualified solder joints corresponding to the welding position can be obtained. The first feature set is matched with the first reference feature set to obtain the first matching value, and welding detection is initially realized from the appearance level. Secondly, when the first matching value is within the first preset range, it indicates that the welding appearance is average. The first welding parameters and the first welding detection standard parameters corresponding to the first welding position can be determined through the welding system. In this way, the welding parameters corresponding to the welding position and the corresponding welding detection standard parameters can be obtained. Thirdly, not only can the multi-scale decomposition algorithm corresponding to the welding process be adopted, so as to ensure the in-depth extraction of the welding features corresponding to the welding process. In addition, the corresponding algorithm control is also based on the welding detection standard, which helps to in-depth extract the welding features corresponding to the welding process. Thus, it helps to improve the welding detection efficiency in the automobile processing process. Fourthly, the low-frequency component part reflects the main body of the image, and the m high-frequency component parts reflect the details of the image. The low-frequency component part and the m high-frequency component parts reflect the influence degree of the main body of the image on the image details. The first welding detection standard parameter reflects the welding detection standard corresponding to the welding process, that is, the corresponding high-frequency feature parameters can be extracted based on the influence degree of the main body of the image on the image details (the characteristics of the image itself) and the welding detection standard corresponding to the welding process. Thus, the welding features corresponding to the welding process can be in-depth extracted. Fifthly, the welding quality is deeply evaluated based on the welding features at the appearance level and the deep welding details, that is, the target matching value is determined according to the first matching value and the second matching value. When the target matching value is greater than or equal to the first upper limit threshold, it indicates that the welding is good, that is, it is determined that the welding detection at the first welding position is qualified. Thus, it helps to improve the welding detection efficiency in the automobile processing process.

[0115] Consistently with the above embodiments, please refer to Figure 2 , Figure 2 FIG. is a schematic structural diagram of an industrial robot provided by an embodiment of the present application. The industrial robot includes a processor, a memory, a communication interface, and one or more programs. Among them, the above one or more programs are stored in the above memory and are configured to be executed by the above processor. In the embodiment of the present application, the industrial robot includes a vision system and a welding system. The above program includes instructions for performing the following steps:

[0116] Obtain a first image of the first welding position through the vision system, where the first welding position is the welded position between the first automotive part and the second automotive part;

[0117] Perform region segmentation on the first image to obtain a first welding area image, perform first feature extraction based on the first welding area image to obtain a first feature set, obtain a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value;

[0118] When the first matching value is within a first preset range, determine, through the welding system, a first welding parameter and first welding detection standard parameters corresponding to the first welding position. The first welding detection standard parameters include k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper limit threshold and a first lower limit threshold;

[0119] Determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameters;

[0120] Perform multi-scale decomposition on the first welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer;

[0121] Perform feature extraction on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameters to obtain k high-frequency feature parameters;

[0122] Match the k high-frequency feature parameters with the k standard high-frequency feature parameters to obtain a second matching value;

[0123] Determine a target matching value according to the first matching value and the second matching value;

[0124] When the target matching value is greater than or equal to the first upper limit threshold, determine that the welding detection of the first welding position is qualified.

[0125] Optionally, the above program further includes instructions for performing the following steps:

[0126] When the first matching value is greater than the first upper limit threshold, determine that the welding detection of the first welding position is qualified;

[0127] Or,

[0128] When the first matching value is less than the first lower limit threshold, determine that the welding detection of the first welding position is unqualified;

[0129] Or,

[0130] When the target matching value is less than the first upper threshold value, it is determined that the welding detection at the first welding position is unqualified.

[0131] Optionally, the first welding detection standard parameter further includes a standard matching value; in terms of determining the target matching value according to the first matching value and the second matching value, the above program includes instructions for performing the following steps:

[0132] Determine a first relative deviation degree between the first upper threshold value and the first matching value;

[0133] When the first relative deviation degree is less than or equal to a preset relative deviation degree, determine a first deviation degree between the second matching value and the standard matching value;

[0134] Determine a first adjustment parameter corresponding to the first deviation degree;

[0135] Determine the target matching value according to the first adjustment parameter and the first matching value;

[0136] When the first relative deviation degree is greater than the preset relative deviation degree, determine a second relative deviation degree between the first matching value and a first lower threshold value;

[0137] Determine a target weight pair corresponding to the second relative deviation degree, the target weight pair includes a first weight and a second weight, and the sum of the first weight and the second weight is 1;

[0138] Perform a weighted operation according to the first matching value, the second matching value and the target weight pair to obtain the target matching value.

[0139] Optionally, in terms of extracting k kinds of high-frequency feature parameters from the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter, the above program includes instructions for performing the following steps:

[0140] Determine a first information entropy of the low-frequency component part;

[0141] Determine a second information entropy of the first welding area image;

[0142] Determine a ratio between the first information entropy and the second information entropy to obtain a first ratio;

[0143] When the first ratio is within a second preset range, extract k kinds of high-frequency feature parameters from the m high-frequency component parts according to the first welding detection standard parameter; the second preset range includes a second upper threshold value and a second lower threshold value.

[0144] Optionally, the above program further includes instructions for performing the following steps:

[0145] When the first ratio is greater than the second upper threshold, determine a third relative deviation between the first ratio and the second upper threshold, determine a first high-frequency image enhancement algorithm corresponding to the third relative deviation, determine a first mean value between the first matching value and the second matching value; determine a first control parameter corresponding to the first high-frequency image enhancement algorithm corresponding to the first mean value;

[0146] Perform image enhancement on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter;

[0147] Extract features from the m high-frequency component parts after image enhancement according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0148] Optionally, the above program further includes instructions for performing the following steps:

[0149] When the first ratio is less than the second lower threshold, determine a fourth relative deviation between the second upper threshold and the first ratio, determine a first low-frequency image enhancement algorithm corresponding to the fourth relative deviation, determine a first difference between the first matching value and the first lower threshold, determine a second control parameter of the first low-frequency image enhancement algorithm corresponding to the first difference;

[0150] Perform image enhancement on the low-frequency component part according to the first low-frequency image enhancement algorithm and the second control parameter;

[0151] Reconstruct the second welding area image according to the low-frequency component part and the m high-frequency component parts after image enhancement;

[0152] Perform multi-scale decomposition on the second welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain m target high-frequency component parts;

[0153] Extract features from the m target high-frequency component parts according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0154] It can be seen that the industrial robot described in the embodiments of the present application includes a vision system and a welding system. The vision system is used to obtain a first image of a first welding position, where the first welding position is the welded position between a first automotive component and a second automotive component; perform region segmentation on the first image to obtain a first welding region image, extract first features according to the first welding region image to obtain a first feature set, obtain a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value; when the first matching value is within a first preset range, determine a first welding parameter and a first welding detection standard parameter corresponding to the first welding position through the welding system, where the first welding detection standard parameter includes k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter; perform multi-scale decomposition on the first welding region image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; extract k high-frequency feature parameters from the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter; match the k high-frequency feature parameters with the k standard high-frequency feature parameters to obtain a second matching value; determine a target matching value according to the first matching value and the second matching value;When the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding detection at the first welding position is qualified. Firstly, the characteristics of the qualified solder joints corresponding to the welding position can be obtained, the first feature set is matched with the first reference feature set to obtain the first matching value, and welding detection is initially realized from the appearance level. Secondly, when the first matching value is within the first preset range, it indicates that the welding appearance is average. The first welding parameters and the first welding detection standard parameters corresponding to the first welding position can be determined through the welding system. In this way, the welding parameters corresponding to the welding position and the corresponding welding detection standard parameters can be obtained. Thirdly, not only can a multi-scale decomposition algorithm corresponding to the welding process be adopted, which can ensure the in-depth extraction of welding features corresponding to the welding process. Additionally, algorithm control is also performed based on the welding detection standard, which helps to in-depth extract the welding features corresponding to the welding process. Thus, it helps to improve the welding detection efficiency in the automotive processing process. Fourthly, the low-frequency component part reflects the main body of the image, and m high-frequency component parts reflect the details of the image. The low-frequency component part and the m high-frequency component parts reflect the influence degree of the main body of the image on the image details. The first welding detection standard parameter reflects the welding detection standard corresponding to the welding process, that is, the corresponding high-frequency feature parameters can be extracted based on the influence degree of the main body of the image on the image details (the characteristics of the image itself) and the welding detection standard corresponding to the welding process. Thus, the welding features corresponding to the welding process can be in-depth extracted. Fifthly, the welding quality is deeply evaluated based on the welding features at the appearance level and the deep welding details, that is, the target matching value is determined according to the first matching value and the second matching value. When the target matching value is greater than or equal to the first upper limit threshold, it indicates that the welding is good, that is, it is determined that the welding detection at the first welding position is qualified. Thus, it helps to improve the welding detection efficiency in the automotive processing process.

[0155] Figure 3 It is a functional unit composition block diagram of a welding detection system 300 in the automotive processing process involved in the embodiments of the present application. The welding detection system 300 in the automotive processing process is applied to an industrial robot. The industrial robot includes a vision system and a welding system. The welding detection system 300 in the automotive processing process includes:

[0156] An acquisition unit 301, configured to acquire a first image of a first welding position through the vision system, where the first welding position is a welded position between a first automotive component and a second automotive component;

[0157] A segmentation unit 302, configured to perform region segmentation on the first image to obtain a first welding region image, perform first feature extraction according to the first welding region image to obtain a first feature set, acquire a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value;

[0158] A determination unit 303, configured to, when the first matching value is within a first preset range, determine, through the welding system, a first welding parameter and a first welding detection standard parameter corresponding to the first welding position. The first welding detection standard parameter includes k standard high-frequency characteristic parameters, and each standard high-frequency characteristic parameter corresponds to at least one high-frequency characteristic type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper limit threshold and a first lower limit threshold; determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter;

[0159] A decomposition unit 304, configured to perform multi-scale decomposition on the first welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer;

[0160] An extraction unit 305, configured to perform feature extraction on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter to obtain k high-frequency characteristic parameters;

[0161] A matching unit 306, configured to match the k high-frequency characteristic parameters with the k standard high-frequency characteristic parameters to obtain a second matching value;

[0162] The determination unit 303 is configured to determine a target matching value according to the first matching value and the second matching value; when the target matching value is greater than or equal to the first upper limit threshold, determine that the welding detection of the first welding position is qualified.

[0163] Optionally, the welding detection system 300 in the automobile processing process is further specifically configured to:

[0164] When the first matching value is greater than the first upper limit threshold, determine that the welding detection of the first welding position is qualified;

[0165] Or,

[0166] When the first matching value is less than the first lower limit threshold, determine that the welding detection of the first welding position is unqualified;

[0167] Or,

[0168] When the target matching value is less than the first upper limit threshold, determine that the welding detection of the first welding position is unqualified.

[0169] Optionally, the first welding detection standard parameter further includes a standard matching value; in terms of determining the target matching value according to the first matching value and the second matching value, the determining unit 303 is specifically configured to:

[0170] Determine a first relative deviation degree between the first upper threshold and the first matching value;

[0171] When the first relative deviation degree is less than or equal to a preset relative deviation degree, determine a first deviation degree between the second matching value and the standard matching value;

[0172] Determine a first adjustment parameter corresponding to the first deviation degree;

[0173] Determine the target matching value according to the first adjustment parameter and the first matching value;

[0174] When the first relative deviation degree is greater than the preset relative deviation degree, determine a second relative deviation degree between the first matching value and the first lower threshold;

[0175] Determine a target weight pair corresponding to the second relative deviation degree, the target weight pair includes a first weight and a second weight, and the sum of the first weight and the second weight is 1;

[0176] Perform a weighted operation according to the first matching value, the second matching value, and the target weight pair to obtain the target matching value.

[0177] Optionally, in terms of extracting k high-frequency characteristic parameters from the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter, the extraction unit 305 is specifically configured to:

[0178] Determine a first information entropy of the low-frequency component part;

[0179] Determine a second information entropy of the first welding area image;

[0180] Determine a ratio between the first information entropy and the second information entropy to obtain a first ratio;

[0181] When the first ratio is within a second preset range, extract k high-frequency characteristic parameters from the m high-frequency component parts according to the first welding detection standard parameter; the second preset range includes a second upper threshold and a second lower threshold.

[0182] Optionally, the welding detection system 300 during the automobile processing process is further specifically configured to:

[0183] When the first ratio is greater than the second upper threshold, determine a third relative deviation between the first ratio and the second upper threshold, determine a first high-frequency image enhancement algorithm corresponding to the third relative deviation, and determine a first mean value between the first matching value and the second matching value; determine a first control parameter corresponding to the first high-frequency image enhancement algorithm corresponding to the first mean value;

[0184] Perform image enhancement on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter;

[0185] Extract features from the m high-frequency component parts after image enhancement according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0186] Optionally, the welding detection system 300 in the automobile processing process is further specifically configured to:

[0187] When the first ratio is less than the second lower threshold, determine a fourth relative deviation between the second upper threshold and the first ratio, determine a first low-frequency image enhancement algorithm corresponding to the fourth relative deviation, determine a first difference between the first matching value and the first lower threshold, and determine a second control parameter of the first low-frequency image enhancement algorithm corresponding to the first difference;

[0188] Perform image enhancement on the low-frequency component part according to the first low-frequency image enhancement algorithm and the second control parameter;

[0189] Reconstruct the second welding area image according to the image-enhanced low-frequency component part and the m high-frequency component parts;

[0190] Perform multi-scale decomposition on the second welding area image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain m target high-frequency component parts;

[0191] Extract features from the m target high-frequency component parts according to the first welding detection standard parameter to obtain the k high-frequency feature parameters.

[0192] It can be seen that the welding detection system in the automobile processing process described in the embodiments of the present application is applied to an industrial robot, which includes a vision system and a welding system. The vision system is used to obtain a first image of a first welding position, where the first welding position is the welded position between a first automotive component and a second automotive component. The first image is subjected to region segmentation to obtain a first welding region image, and first feature extraction is performed on the first welding region image to obtain a first feature set. A first reference feature set corresponding to the first welding position is obtained, and the first feature set is matched with the first reference feature set to obtain a first matching value. When the first matching value is within a first preset range, the welding system is used to determine a first welding parameter and a first welding detection standard parameter corresponding to the first welding position. The first welding detection standard parameter includes k standard high-frequency feature parameters, and each standard high-frequency feature parameter corresponds to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter are determined; the first welding region image is subjected to multi-scale decomposition according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; feature extraction is performed on the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts, and the first welding detection standard parameter to obtain k high-frequency feature parameters; the k high-frequency feature parameters are matched with the k standard high-frequency feature parameters to obtain a second matching value; a target matching value is determined according to the first matching value and the second matching value;When the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding inspection at the first welding position is qualified. Firstly, the characteristics of qualified solder joints corresponding to the welding position can be obtained, and the first feature set is matched with the first reference feature set to obtain the first matching value, thus realizing welding inspection preliminarily from the appearance level. Secondly, when the first matching value is within the first preset range, it indicates that the welding appearance is average. The first welding parameters and the first welding inspection standard parameters corresponding to the first welding position can be determined through the welding system. In this way, the welding parameters corresponding to the welding position and the corresponding welding inspection standard parameters can be obtained. Thirdly, not only can the multi-scale decomposition algorithm corresponding to the welding process be adopted, so as to ensure the in-depth extraction of welding features corresponding to the welding process. In addition, corresponding algorithm control is also carried out based on the welding inspection standard, which helps to in-depth extract welding features corresponding to the welding process. Therefore, it helps to improve the welding inspection efficiency in the automotive processing process. Fourthly, the low-frequency component part reflects the main body of the image, and m high-frequency component parts reflect the details of the image. The low-frequency component part and the m high-frequency component parts reflect the influence degree of the image main body on the image details, and the first welding inspection standard parameter reflects the welding inspection standard corresponding to the welding process, that is, the corresponding high-frequency feature parameters can be extracted based on the influence degree of the image main body on the image details (the characteristics of the image itself) and the welding inspection standard corresponding to the welding process, so as to in-depth extract welding features corresponding to the welding process. Fifthly, the welding quality is deeply evaluated based on the welding features at the appearance level and the deep welding details, that is, the target matching value is determined according to the first matching value and the second matching value. When the target matching value is greater than or equal to the first upper limit threshold, it indicates that the welding is good, that is, it is determined that the welding inspection at the first welding position is qualified. Therefore, it helps to improve the welding inspection efficiency in the automotive processing process.

[0193] The embodiment of the present application also provides a computer storage medium. Among them, this computer storage medium stores a computer program for electronic data exchange, and this computer program enables the computer to execute part or all of the steps of any method recorded in the above method embodiments. The above computer includes an industrial robot.

[0194] The embodiment of the present application also provides a computer program product. The above computer program product includes a non-transitory computer-readable storage medium storing a computer program, and the above computer program is operable to enable the computer to execute part or all of the steps of any method recorded in the above method embodiments. This computer program product can be a software installation package, and the above computer includes an industrial robot.

[0195] It should be noted that, for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0196] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0197] In several embodiments provided by this application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above division of units is only a logical function division. In actual implementation, there can be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the devices or units can be in electrical or other forms.

[0198] The units described as separate components above may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0199] In addition, in each embodiment of this application, the functional units can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0200] If the above integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the above methods in various embodiments of this application. The aforementioned memory includes: various media such as USB flash drives, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), external hard drives, magnetic disks, or optical discs that can store program codes.

[0201] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. This program can be stored in a computer-readable memory, and the memory can include: flash drives, read-only memories (English: Read-Only Memory, abbreviated as: ROM), random access memories (English: Random Access Memory, abbreviated as: RAM), magnetic disks, or optical discs, etc.

[0202] The above has introduced the embodiments of this application in detail. Specific examples are used in this article to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A welding detection method in an automobile processing process, characterized in that: Applied to an industrial robot, the industrial robot includes a vision system and a welding system, and the method includes: Acquire a first image for a first welding position by the visual system, where the first welding position is a welded position between a first automobile component and a second automobile component; Performing region segmentation on the first image to obtain a first welding region image, performing first feature extraction based on the first welding region image to obtain a first feature set, acquiring a first reference feature set corresponding to the first welding position, and matching the first feature set with the first reference feature set to obtain a first matching value; When the first matching value is within a first preset range, the first welding parameter and the first welding detection standard parameter corresponding to the first welding position are determined by the welding system, the first welding detection standard parameter includes k standard high-frequency characteristic parameters, each standard high-frequency characteristic parameter corresponds to at least one high-frequency characteristic type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; determining a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter; The first welding area image is multi-scale decomposed according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; Extracting features of the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter to obtain k kinds of high-frequency feature parameters; Matching the k kinds of high-frequency characteristic parameters with the k kinds of standard high-frequency characteristic parameters to obtain a second matching value; Determine a target matching value according to the first matching value and the second matching value; When the target matching value is greater than the first upper limit threshold, determining that the welding inspection of the first welding position is qualified; The feature extraction of the m high-frequency component parts is performed according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter to obtain k kinds of high-frequency feature parameters, including: Determining a first information entropy of the low-frequency component part; determining a second information entropy of the first welding area image; Determine a ratio between the first information entropy and the second information entropy to obtain a first ratio; When the first ratio is within a second preset range, feature extraction is performed on the m high-frequency component parts according to the first welding detection standard parameter to obtain the k high-frequency feature parameters; the second preset range includes a second upper threshold and a second lower threshold; Wherein, the method further comprises: When the first ratio is greater than the second upper limit threshold, determining a third relative deviation between the first ratio and the second upper limit threshold, determining a first high-frequency image enhancement algorithm corresponding to the third relative deviation, determining a first mean between the first matching value and the second matching value; and determining a first control parameter corresponding to the first high-frequency image enhancement algorithm corresponding to the first mean; Performing image enhancement on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter; The m high-frequency component parts after image enhancement are subjected to feature extraction according to the first welding detection standard parameters to obtain the k kinds of high-frequency feature parameters.

2. The method according to claim 1, characterized in that: The method further comprises: When the first matching value is greater than the first upper limit threshold, it is determined that the welding detection of the first welding position is qualified; or, When the first matching value is less than the first lower limit threshold, determining that the welding detection of the first welding position is unqualified; or, When the target matching value is less than the first upper limit threshold, it is determined that the welding inspection of the first welding position is unqualified.

3. The method according to claim 1 or 2, characterized in that: The first welding detection standard parameter also includes a standard matching value; and determining a target matching value according to the first matching value and the second matching value includes: Determining a first relative deviation between the first upper threshold and the first matching value; When the first relative deviation is less than or equal to a preset relative deviation, determining a first deviation between the second matching value and the standard matching value; determining a first adjustment parameter corresponding to the first deviation; Determining the target matching value according to the first adjustment parameter and the first matching value; When the first relative deviation is greater than the preset relative deviation, determining a second relative deviation between the first matching value and the first lower threshold; Determine a target weight pair corresponding to the second relative deviation, the target weight pair comprising a first weight and a second weight, and a sum of the first weight and the second weight is 1; A weighted operation is performed according to the first matching value, the second matching value and the target weight to obtain the target matching value.

4. A welding detection system in the automobile processing process, characterized in that: Applied to an industrial robot, the industrial robot includes a visual system and a welding system, and the welding detection system in the automobile processing process includes: An acquisition unit, configured to acquire a first image for a first welding position through the visual system, where the first welding position is a welded position between a first automobile component and a second automobile component; a segmentation unit, configured to perform region segmentation on the first image to obtain a first welding region image, perform first feature extraction based on the first welding region image to obtain a first feature set, obtain a first reference feature set corresponding to the first welding position, and match the first feature set with the first reference feature set to obtain a first matching value; a determination unit, configured to determine, by the welding system, a first welding parameter and a first welding detection standard parameter corresponding to the first welding position when the first matching value is within a first preset range, the first welding detection standard parameter including k standard high-frequency feature parameters, each standard high-frequency feature parameter corresponding to at least one high-frequency feature type of at least one high-frequency component part; k is a positive integer; the first preset range includes a first upper threshold and a first lower threshold; determine a first multi-scale decomposition algorithm corresponding to the first welding parameter and a first algorithm control parameter of the first multi-scale decomposition algorithm corresponding to the first welding detection standard parameter; a decomposition unit, configured to perform multi-scale decomposition on the first welding region image according to the first multi-scale decomposition algorithm and the first algorithm control parameter to obtain a low-frequency component part and m high-frequency component parts; m is a positive integer; an extraction unit, configured to extract features of the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter, so as to obtain k kinds of high-frequency feature parameters; A matching unit, configured to match the k kinds of high-frequency characteristic parameters with the k kinds of standard high-frequency characteristic parameters to obtain a second matching value; The determining unit is used to determine a target matching value according to the first matching value and the second matching value; when the target matching value is greater than or equal to the first upper limit threshold, it is determined that the welding detection of the first welding position is qualified; Wherein, in the aspect of extracting features of the m high-frequency component parts according to the low-frequency component part, the m high-frequency component parts and the first welding detection standard parameter to obtain k kinds of high-frequency feature parameters, the extraction unit is specifically used for: Determining a first information entropy of the low-frequency component part; determining a second information entropy of the first welding area image; Determine a ratio between the first information entropy and the second information entropy to obtain a first ratio; When the first ratio is within a second preset range, feature extraction is performed on the m high-frequency component parts according to the first welding detection standard parameter to obtain the k high-frequency feature parameters; the second preset range includes a second upper threshold and a second lower threshold; The welding detection system in the automobile processing process is also specifically used for: When the first ratio is greater than the second upper limit threshold, determining a third relative deviation between the first ratio and the second upper limit threshold, determining a first high-frequency image enhancement algorithm corresponding to the third relative deviation, determining a first mean between the first matching value and the second matching value; and determining a first control parameter corresponding to the first high-frequency image enhancement algorithm corresponding to the first mean; Performing image enhancement on the m high-frequency component parts according to the first high-frequency image enhancement algorithm and the first control parameter; The m high-frequency component parts after image enhancement are subjected to feature extraction according to the first welding detection standard parameters to obtain the k kinds of high-frequency feature parameters.

5. The welding detection system in the automobile processing process according to claim 4 is characterized in that: The welding detection system in the automobile processing process is also specifically used for: When the first matching value is greater than the first upper limit threshold, it is determined that the welding detection of the first welding position is qualified; or, When the first matching value is less than the first lower limit threshold, determining that the welding detection of the first welding position is unqualified; or, When the target matching value is less than the first upper limit threshold, it is determined that the welding inspection of the first welding position is unqualified.

6. The welding detection system in the automobile processing process according to claim 4 or 5, characterized in that: The first welding detection standard parameter also includes a standard matching value; in determining the target matching value according to the first matching value and the second matching value, the determining unit is specifically used to: Determining a first relative deviation between the first upper threshold and the first matching value; When the first relative deviation is less than or equal to a preset relative deviation, determining a first deviation between the second matching value and the standard matching value; determining a first adjustment parameter corresponding to the first deviation; Determining the target matching value according to the first adjustment parameter and the first matching value; When the first relative deviation is greater than the preset relative deviation, determining a second relative deviation between the first matching value and the first lower threshold; Determine a target weight pair corresponding to the second relative deviation, the target weight pair comprising a first weight and a second weight, and a sum of the first weight and the second weight is 1; A weighted operation is performed according to the first matching value, the second matching value and the target weight to obtain the target matching value.

Images