Automatic visual identification and hardness detection point location automatic planning method for inlaying and welding line materials

Abstract

The invention relates to the field of detection equipment, and discloses an automatic visual identification and hardness detection point position automatic planning method for an inlaid and welded joint material, which comprises an image acquisition module, the image acquisition module comprises an industrial camera, and the image acquisition module shoots the top surface of the inlaid and welded joint material so as to realize an image acquisition effect. And the image acquisition module is connected with an AI image recognition module, and the AI image recognition module analyzes feature information in the image by using an AI intelligent recognition technology, so that the accuracy of subsequent image judgment of the detection module is improved. According to the invention, tedious edge scanning and image splicing operations in traditional hardness detection are abandoned, an industrial camera in the image acquisition module is used for shooting once to obtain top surface images of mosaic and weld materials, the image acquisition time is greatly shortened, the image feature extraction module is used for rapidly extracting colors, textures and shapes in the images, and the detection accuracy is improved. And the subsequent detection efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of detection equipment, and in particular to a method for automatic visual recognition of inlay and weld materials and automatic planning of hardness detection points. Background Art

[0002] In the field of material testing, hardness testing of mounted specimens is a key step in evaluating material performance. In the past, this testing process was usually performed by relying on the built-in magnification lens of the hardness tester. In this way, the edge scanning function of the lens is required to acquire images of the mounted specimens. Since the magnification lens has a limited field of view and often cannot cover the entire specimen at once, image stitching operations are also required to obtain complete specimen image information in preparation for subsequent hardness test point selection and path planning.

[0003] However, this traditional hardness testing method has many disadvantages. First of all, the entire operation process is extremely cumbersome, and it takes a lot of time and effort from edge scanning to image stitching. The operator needs to carefully operate the magnification lens of the hardness tester to ensure that the scanning process is complete and accurate. Image stitching also requires a high degree of attention to ensure the continuity and accuracy of the image. Secondly, after the image stitching is completed, the selection of hardness testing points and path planning are completely dependent on manual operation. Manual selection of hardness testing points is prone to errors and is affected by subjective factors of the operator, such as visual fatigue and operating proficiency, which may lead to unreasonable selection of points or omission of key testing points. Moreover, manually planned testing paths often lack scientificity and efficiency, and cannot fully consider the optimal solution for testing efficiency and equipment movement, further increasing testing time and cost. Summary of the invention

[0004] In view of the shortcomings of the prior art, the present invention provides a method for automatic visual recognition and automatic planning of hardness detection points for inlay and weld materials, which solves the many drawbacks of traditional inlay sample hardness detection methods and the problem that the operation process from edge scanning to image stitching is extremely cumbersome, time-consuming and labor-intensive.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials, including an image acquisition module, the image acquisition module includes an industrial camera, the image acquisition module shoots the top surface of the inlay and weld materials, thereby achieving the image acquisition effect, the image acquisition module is connected to an AI image recognition module, the AI ​​image recognition module uses AI intelligent recognition technology to analyze the feature information in the image, and improves the accuracy of the subsequent image judgment of the detection module, the AI ​​image recognition module is connected to a detection module, the detection module compares the welding points and paths, so as to determine whether the inlay and weld materials meet the production standards, and sends the conclusion to the feedback conclusion module, the detection module is connected to a feedback conclusion module, the feedback conclusion module receives the conclusion given by the detection module, and reminds the staff of the result.

[0006] Preferably, the image acquisition module includes a camera module, which photographs the top of the material below and transmits the acquired and photographed image to an image feature extraction module. The camera module is connected to the image feature extraction module, and the image feature extraction module distinguishes different weld materials according to various features in the image, or identifies the type of material according to the texture features of the material.

[0007] Preferably, the features in the image feature extraction module include: color, texture, and shape.

[0008] Preferably, the AI ​​image recognition module includes a neural network model training module, which is the basis of the entire AI image recognition. It uses a large amount of pre-collected and annotated inlay and weld material image data for training. During the training process, the neural network continuously adjusts its internal parameters and learns to accurately identify various features in the image and the corresponding material types. The content of the AI ​​recognition includes but is not limited to: weld material: heat affected zone / parent material / weld metal / surface / core / upper part of weld / bottom of weld / root weld / last weld; inlay material: inlay material / sample material.

[0009] Preferably, the neural network model training module is connected to a module for automatically distinguishing between inlays and materials. The module for automatically distinguishing between inlays and materials relies on the data trained by the neural network model training module to accurately identify which parts in the image are inlays and which parts are materials that need to be tested for hardness. Even when the appearance of the material and the inlay are similar or there are interference factors, the distinction can be made, thereby providing an accurate target area for subsequent detection planning.

[0010] Preferably, the detection module includes a detection requirement input module, which allows the operator to input specific requirements for hardness testing, including the distribution density of detection points and detection accuracy requirements. The detection requirement input module is connected to a point planning module, which automatically plans hardness detection points in the material area according to the recognition results of the automatic distinction between inlay materials and material modules and the settings of the detection requirement input module, and reasonably determines the positions of the detection points as required, taking into account the shape, size and special parts of the material.

[0011] Preferably, in the point planning module, the positions of the detection points are reasonably determined as follows: for large-area materials, the detection points are evenly distributed according to the set density; for key parts, including the junctions of welds and mosaics, additional detection points are added to ensure that these areas prone to problems are focused on inspection.

[0012] Preferably, the point planning module is connected to a path planning module, and the path planning module is responsible for planning the moving path of the detection equipment between various detection points. It will comprehensively consider the detection efficiency, the feasibility of equipment movement and safety factors, and adopt an optimized path planning algorithm to enable the detection equipment to traverse all detection points in the shortest time and in the most reasonable order, thereby improving the detection efficiency. At the same time, planning standards are set in the point planning module and the path planning module, wherein the specific planning standards are: weld: built-in weld detection standard detection points within the GB_T_2654-2008 / GBT31032-2023 / ISO_209015-1-2011_20E standard, and also supports user-defined points such as matrix points; inlay material: supports automatic built-in paths such as along the edge, along the tooth, along the normal, and also supports user-defined points; the path planning module is connected to a comparison module, and the comparison module implements the detection plan according to the conclusions drawn by the path planning module and the point planning module, monitors the image content, and determines whether the welding material meets the production standards.

[0013] Preferably, the specific algorithm and calculation formula of the path planning module are:

[0014] 1. Calculation of the number of detection points:

[0015] n=ρ×L×W

[0016] Where: n represents the number of detection points, ρ represents the distribution density of detection points, L represents the length of the detection object, and W represents the width of the detection object;

[0017] 2. Formula for increasing key parts detection points:

[0018] n extra =k1×l weld +k2×wweld

[0019] Where: n extra represents the number of additional inspection points at the joint; k1 and k2 represent coefficients determined based on experience or experiments, which reflect the influence of weld length and width on the number of inspection points; l weld Represents the length of the joint between the weld and the mosaic, w weld Represents the width of the weld where the weld meets the mosaic.

[0020] A method for using a system for automatic visual recognition and hardness detection point automatic planning of inlay and weld materials, characterized in that it comprises the following steps:

[0021] S1. Image acquisition stage: The top surface of the inlay and weld materials is photographed using the industrial camera in the image acquisition module. The camera module obtains the top image of the material and transmits it to the image feature extraction module. By collecting the color, texture, and shape features of the image, raw data is provided for subsequent processing;

[0022] S2. Image recognition stage based on AI neural network: The neural network model training module uses a large amount of labeled inlay and weld material image data to train the neural network, adjusts the internal parameters to learn the correspondence between image features and material types, and then the automatic inlay and material distinction module relies on the trained data to use the feature vector-based classification algorithm, threshold-based distinction method and clustering algorithm to accurately identify the material part of the image that needs to be tested for hardness, providing an accurate target area for detection planning;

[0023] S3. Automatic planning of detection points and paths: The detection requirement input module receives the operator's input including the detection point distribution density, detection accuracy requirements and specific requirements for hardness testing. The point planning module automatically distinguishes the results and input requirements, evenly distributes the detection points according to the density for large-area materials, and increases the detection points for the key parts of the joint between the weld and the mosaic according to the key parts detection point increase formula; the path planning module calculates and determines the number of points based on the number of detection points, and comprehensively considers the detection efficiency, equipment movement feasibility and safety factors, and uses the optimization algorithm to plan the movement path of the detection equipment between the detection points;

[0024] S4. Detection and judgment stage: The comparison module implements detection according to the plan obtained by the path planning module and the point planning module. During the detection process, the image content is monitored in real time to determine whether the welding material meets the production standards, thereby completing the automatic visual recognition of the inlay and weld materials and the automatic planning and detection of the hardness detection points.

[0025] The present invention provides a method for automatic visual recognition of inlay and weld materials and automatic planning of hardness detection points. It has the following beneficial effects:

[0026] 1. The present invention abandons the cumbersome edge scanning and image stitching operations in traditional hardness testing, and uses the industrial camera in the image acquisition module to obtain the top surface image of the inlay and weld material in one shot, which greatly shortens the image acquisition time. At the same time, the image feature extraction module is used to quickly extract the color, texture, and shape in the image, thereby improving the subsequent detection efficiency.

[0027] 2. The AI ​​image recognition module of the present invention uses a large amount of pre-collected and annotated image data to train the neural network, and can accurately identify various features in the image and the corresponding material types. The automatic differentiation module for inlays and materials is based on the training results, and can reliably distinguish between materials and inlays even when the materials and inlays are similar in appearance or there are interference factors, providing an accurate target area for subsequent detection, avoiding the problem of detection point errors caused by manual recognition errors.

[0028] 3. The present invention sets a point planning module and a path planning module, and uses formula calculation to comprehensively consider multiple factors such as detection efficiency, feasibility and safety of equipment movement, and uses an optimized path planning algorithm. The number of points is determined by relevant formulas such as the number of detection points, and then the optimal path is planned, so that the detection equipment can traverse all detection points in the shortest time and the most reasonable order. This not only reduces the invalid movement time of the equipment and improves the detection efficiency, but also reduces the risk of equipment loss caused by unreasonable paths, and ensures the fluency and stability of the detection work. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 is a system flow chart of the present invention;

[0030] Figure 2 This is a flow chart of the AI ​​image recognition module of the present invention;

[0031] Figure 3 It is a flow chart of the detection module of the present invention;

[0032] Figure 4 Flow chart of the method of the present invention DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the specification of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] Example:

[0035] Please see attached Figure 1 - Attachment Figure 4 An embodiment of the present invention provides an automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials, including an image acquisition module, the image acquisition module including an industrial camera, the image acquisition module photographs the top surface of the inlay and weld materials, thereby achieving the effect of image acquisition, the image acquisition module is connected to an AI image recognition module, the AI ​​image recognition module uses AI intelligent recognition technology to analyze the feature information in the image, and improves the accuracy of the subsequent image judgment of the detection module, the AI ​​image recognition module is connected to a detection module, the detection module compares the welding points and paths, thereby determining whether the inlay and weld materials meet the production standards, and sends the conclusion to a feedback conclusion module, the detection module is connected to a feedback conclusion module, the feedback conclusion module receives the conclusion given by the detection module, and reminds the staff of the result.

[0036] The image acquisition module includes a camera module, which photographs the top of the material below and transmits the acquired and photographed image to the image feature extraction module. The camera module is connected to the image feature extraction module, and the image feature extraction module distinguishes different weld materials according to various features in the image, or identifies the type of material according to the texture features of the material, and the various features include color, texture, and shape.

[0037] The AI ​​image recognition module includes a neural network model training module, which is the basis of the entire AI image recognition. It uses a large amount of pre-collected and annotated inlay and weld material image data for training. During the training process, the neural network continuously adjusts its internal parameters and learns to accurately identify various features in the image and the corresponding material types. The content of the AI ​​recognition includes but is not limited to: weld material: heat affected zone / parent material / weld metal / surface / core / upper weld / bottom weld / root weld / last weld; inlay material: inlay material / sample material;

[0038] The neural network model training module is connected to a module for automatically distinguishing between inlays and materials. The module for automatically distinguishing between inlays and materials relies on the data trained by the neural network model training module to accurately identify which parts in the image are inlays and which parts are materials that need to be tested for hardness. Even when the appearance of the material and the inlay are similar or there are interference factors, the distinction can be made, thereby providing an accurate target area for subsequent detection planning.

[0039] The algorithm in the automatic differentiation module between inlays and materials:

[0040] 1. Classification algorithm based on feature vector:

[0041] After the neural network model is trained, the feature vector of the image is extracted. For a new image region, assuming its feature vector is x, the goal of SVM is to find a hyperplane w·x+b=0 (for the linearly separable case) to separate data of different categories. The hyperplane parameters w and b are determined by solving the following optimization problem;

[0042]

[0043] Among them, x i and i It is the sample feature vector and the corresponding category label in the training data. In the case of nonlinear separability, the kernel function is used to map the data to a high-dimensional space, and then a hyperplane is found in the high-dimensional space. For the distinction between inlays and materials, different types of inlays and materials are regarded as different categories and classified by SVM.

[0044] 2. Threshold-based differentiation method:

[0045] If it is found through neural network training that certain features (such as the grayscale value of color, the value of texture features, etc.) have obvious threshold differences between inlays and materials, the threshold method is used to distinguish them. For example, assuming that a certain texture feature value extracted from the image is t, after analysis, it is found that when t>T (T is the threshold), it is a material, and when t<T, it is an inlay. The threshold T is obtained through statistical analysis of the training data, such as calculating the distribution of feature values ​​of samples of different categories, and selecting a suitable dividing value as the threshold;

[0046] 3. Based on clustering algorithm:

[0047] For pixels or regions in an image, a K-means clustering algorithm is first used for preliminary classification. The goal of K-means clustering is to divide data points into K clusters so that the sum of the distances from each point to the center of the cluster to which it belongs is minimized. Through clustering, the pixels in the image are grouped according to the similarity of their features (such as color, texture, etc.). In images of inlays and weld materials, the clustering results may roughly separate the pixels of inlays and materials, and then further accurately classify them in combination with a more accurate classification model obtained by neural network training. Assume that the data point x i , the cluster center is μ j , the objective function of K-means clustering is:

[0048]

[0049] The cluster centers are updated iteratively and the clusters to which the data points belong are redistributed until the objective function converges.

[0050] The detection module includes a detection requirement input module, which allows the operator to input specific requirements for hardness testing, including the distribution density of detection points and detection accuracy requirements. The detection requirement input module is connected to a point planning module, which automatically plans hardness detection points in the material area based on the recognition results of the automatic distinction between inlays and materials modules and the settings of the detection requirement input module, and considers the shape, size and special parts of the material to reasonably determine the location of the detection points according to the requirements. For large-area materials, the detection points will be evenly distributed according to the set density; for key parts, including the junction of welds and inlays, additional detection points will be added to ensure that these areas prone to problems are tested in a focused manner;

[0051] The point planning module is connected to a path planning module, which is responsible for planning the moving path of the detection equipment between various detection points. It will comprehensively consider the detection efficiency, the feasibility of the equipment movement and the safety factors, and adopt an optimized path planning algorithm to enable the detection equipment to traverse all the detection points in the shortest time and the most reasonable order, thereby improving the detection efficiency. At the same time, planning standards are set in the point planning module and the path planning module, and the specific planning standards are:

[0052] Weld seam: Built-in weld seam inspection standard detection points in GB_T_2654-2008 / GBT31032-2023 / ISO_209015-1-2011_20E standards, and also supports user-defined points such as matrix points;

[0053] Inlay material: supports automatic along edge, along tooth, along normal and other built-in paths, and also supports user-defined points;

[0054] The specific algorithm and calculation formula of the path planning module are as follows:

[0055] 1. Calculation of the number of detection points:

[0056] n=ρ×L×W

[0057] Where: n represents the number of detection points, ρ represents the distribution density of detection points, L represents the length of the detection object, and W represents the width of the detection object;

[0058] 2. Formula for increasing key parts detection points:

[0059] n extra =k1×l weld +k2×w weld

[0060] Where: n extrarepresents the number of additional inspection points at the joint; k1 and k2 represent coefficients determined based on experience or experiments, which reflect the influence of weld length and width on the number of inspection points; l weld Represents the length of the joint between the weld and the mosaic, w weld Represents the width of the joint between the weld and the mosaic piece;

[0061] The path planning module is connected to a comparison module, which implements a detection plan according to the conclusions drawn by the path planning module and the point planning module, monitors the image content, and determines whether the welding material meets the production standards.

[0062] A method for using an automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials, comprising the following steps:

[0063] S1. Image acquisition stage: The top surface of the inlay and weld materials is photographed using the industrial camera in the image acquisition module. The camera module obtains the top image of the material and transmits it to the image feature extraction module. By collecting the color, texture, and shape features of the image, raw data is provided for subsequent processing;

[0064] S2. Image recognition stage based on AI neural network: The neural network model training module uses a large amount of labeled inlay and weld material image data to train the neural network, adjusts the internal parameters to learn the correspondence between image features and material types, and then the automatic inlay and material distinction module relies on the trained data to use the feature vector-based classification algorithm, threshold-based distinction method and clustering algorithm to accurately identify the material part of the image that needs to be tested for hardness, providing an accurate target area for detection planning;

[0065] S3. Automatic planning of detection points and paths: The detection requirement input module receives the operator's input including the detection point distribution density, detection accuracy requirements and specific requirements for hardness testing. The point planning module automatically distinguishes the results and input requirements, evenly distributes the detection points according to the density for large-area materials, and increases the detection points for the key parts of the joint between the weld and the mosaic according to the key parts detection point increase formula; the path planning module calculates and determines the number of points based on the number of detection points, and comprehensively considers the detection efficiency, equipment movement feasibility and safety factors, and uses the optimization algorithm to plan the movement path of the detection equipment between the detection points;

[0066] S4. Detection and judgment stage: The comparison module implements the detection according to the scheme obtained by the path planning module and the point planning module. During the detection process, the image content is monitored in real time to determine whether the welding material meets the production standards, thereby completing the automatic visual recognition of the inlay and weld materials and the automatic planning and detection of the hardness detection points.

[0067] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials, characterized in that: It includes an image acquisition module, which includes an industrial camera. The image acquisition module shoots the top surface of the inlay and weld materials to achieve the effect of image acquisition. The image acquisition module is connected to an AI image recognition module. The AI ​​image recognition module uses AI intelligent recognition technology to analyze the feature information in the image to improve the accuracy of the subsequent image judgment by the detection module. The AI ​​image recognition module is connected to a detection module. The detection module compares the welding points and paths to determine whether the inlay and weld materials meet the production standards and sends the conclusion to the feedback conclusion module. The detection module is connected to a feedback conclusion module. The feedback conclusion module receives the conclusion given by the detection module and reminds the staff of the result.

2. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 1 is characterized in that: The image acquisition module includes a camera module, which photographs the top of the material below and transmits the acquired and photographed image to the image feature extraction module. The camera module is connected to the image feature extraction module, and the image feature extraction module distinguishes different weld materials according to various features in the image, or identifies the type of material according to the texture features of the material.

3. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 2 is characterized in that: The various features in the image feature extraction module include: color, texture, and shape.

4. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 1 is characterized in that: The AI ​​image recognition module includes a neural network model training module, which is the basis of the entire AI image recognition. It uses a large amount of pre-collected and annotated inlay and weld material image data for training. During the training process, the neural network continuously adjusts its internal parameters and learns to accurately identify various features in the image and the corresponding material types. The content of the AI ​​recognition includes but is not limited to: weld material: heat affected zone / parent material / weld metal / surface / core / upper part of weld / bottom of weld / root weld / last weld; inlay material: inlay material / sample material.

5. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 4 is characterized in that: The neural network model training module is connected to a module for automatically distinguishing between inlays and materials. The module for automatically distinguishing between inlays and materials relies on the data trained by the neural network model training module to accurately identify which parts of the image are inlays and which parts are materials that need to be tested for hardness. Even when the appearance of the material and the inlay are similar or there are interference factors, the distinction can be made, thereby providing an accurate target area for subsequent detection planning.

6. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 1 is characterized in that: The detection module includes a detection requirement input module, which allows the operator to input specific requirements for hardness testing, including the distribution density of detection points and detection accuracy requirements. The detection requirement input module is connected to a point planning module, which automatically plans hardness detection points in the material area according to the recognition results of the automatic distinction between inlay materials and material modules and the settings of the detection requirement input module, taking into account the shape, size and special parts of the material, and reasonably determining the positions of the detection points as required.

7. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 6 is characterized in that: In the point planning module, the positions of the detection points are reasonably determined as follows: for large-area materials, the detection points are evenly distributed according to the set density; for key parts, including the junctions of welds and mosaics, additional detection points are added to ensure that these areas prone to problems are focused on inspection.

8. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 6 is characterized in that: The point planning module is connected to a path planning module, which is responsible for planning the moving path of the detection equipment between various detection points. It will comprehensively consider the detection efficiency, the feasibility of equipment movement and safety factors, and adopt an optimized path planning algorithm to enable the detection equipment to traverse all detection points in the shortest time and the most reasonable order to improve the detection efficiency. At the same time, planning standards are set in the point planning module and the path planning module, and the specific planning standards are: weld: built-in weld detection standard detection points within the GB_T_2654-2008 / GBT31032-2023 / ISO_209015-1-2011_20E standard, and also supports user-defined points such as matrix points; inlay material: supports automatic built-in paths such as along the edge, along the tooth, along the normal, and also supports user-defined points; the path planning module is connected to a comparison module, and the comparison module implements the detection plan according to the conclusions drawn by the path planning module and the point planning module, monitors the image content, and determines whether the welding material meets the production standards.

9. The automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to claim 8, characterized in that: The specific algorithm and calculation formula of the path planning module are as follows:

1. Calculation of the number of detection points: n=ρ×L×W Where: n represents the number of detection points, ρ represents the distribution density of detection points, L represents the length of the detection object, and W represents the width of the detection object; 2. Formula for increasing key parts detection points: n extra =k1×l weld +k2×w weld Where: n extra represents the number of additional inspection points at the joint; k1 and k2 represent coefficients determined based on experience or experiments, which reflect the influence of weld length and width on the number of inspection points; l weld Represents the length of the joint between the weld and the mosaic, w weld Represents the width of the weld where the weld meets the mosaic.

10. A method for using the automatic visual recognition and hardness detection point automatic planning system for inlay and weld materials according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Image acquisition stage: The top surface of the inlay and weld materials is photographed using the industrial camera in the image acquisition module. The camera module obtains the top image of the material and transmits it to the image feature extraction module. By collecting the color, texture, and shape features of the image, raw data is provided for subsequent processing; S2. Image recognition stage based on AI neural network: The neural network model training module uses a large amount of labeled inlay and weld material image data to train the neural network, adjusts the internal parameters to learn the correspondence between image features and material types, and then the automatic inlay and material distinction module relies on the trained data to use the feature vector-based classification algorithm, threshold-based distinction method and clustering algorithm to accurately identify the material part of the image that needs to be tested for hardness, providing an accurate target area for detection planning; S3. Automatic planning of detection points and paths: The detection requirement input module receives the input of the operator, including the detection point distribution density, detection accuracy requirements and specific requirements for hardness testing. The point planning module automatically distinguishes the results and input requirements, evenly distributes the detection points according to the density for large-area materials, and increases the detection points for the key parts of the joint between the weld and the mosaic according to the key parts detection point increase formula; The path planning module calculates the number of detection points based on the number of detection points, and comprehensively considers the detection efficiency, feasibility of equipment movement and safety factors, and uses the optimization algorithm to plan the movement path of the detection equipment between the detection points; S4. Detection and judgment stage: The comparison module implements detection according to the plan obtained by the path planning module and the point planning module. During the detection process, the image content is monitored in real time to determine whether the welding material meets the production standards, thereby completing the automatic visual recognition of the inlay and weld materials and the automatic planning and detection of the hardness detection points.

Images