Device and method for detecting technological parameters of parts

Abstract

The invention discloses a device and method for detecting technological parameters of a part, and relates to the field related to automatic control, and the device comprises a part information obtaining module which is used for carrying out image collection and obtaining part laser point cloud information; the part space model acquisition module is used for performing space registration fusion on the part image information and the part laser point cloud information to obtain a standard part space model; the multi-dimensional loss analysis module is used for performing multi-dimensional loss analysis on the target and standard part space models; the part process detection characteristic parameter acquisition module is used for acquiring internal defect characteristic parameters and part process detection characteristic parameters; and the part production optimization control module is used for carrying out process parameter compensation optimization and carrying out production control through part production process optimization parameters. The technical problems that existing part technological parameter detection is insufficient in detection precision and low in detection efficiency are solved, and the technical effects that the parts are efficiently and accurately detected, and adaptive optimization of the production technology is achieved are achieved.

Description

Technical Field

[0001] The present application relates to the field of automation control, and in particular to a device and method for detecting process parameters of a part. Background Art

[0002] In the manufacturing industry, the detection of process parameters of parts is a key link to ensure product quality and production efficiency. Traditional detection methods, such as using vernier calipers, micrometers and other measuring tools to measure the dimensions of parts, have a certain degree of accuracy, but the measurement range is limited, and it is difficult to accurately measure parts with complex shapes, and it takes a long time, which cannot meet the production efficiency requirements of modern manufacturing.

[0003] In the current related technologies, the process parameter detection of parts has technical problems such as insufficient detection accuracy and low detection efficiency. Summary of the invention

[0004] The present application provides a device and method for detecting process parameters of parts, uses an industrial camera and a three-dimensional scanner to simultaneously collect image information and laser point cloud information of parts, spatially aligns and fuses the image information and laser point cloud information of parts, generates a target part space model, and performs multi-dimensional loss analysis with the standard part space model to obtain part surface deviation parameters, obtains internal defect characteristic parameters of parts through ultrasonic sensor detection, obtains part process detection characteristic parameters in combination with the part surface deviation parameters, performs process parameter compensation optimization based on the part process detection characteristic parameters, obtains part production process optimization parameters, and uses these parameters to perform production optimization control on target parts, thereby achieving the technical effects of efficient and accurate detection of parts, adaptive optimization of production processes, and improving product process quality.

[0005] The present application provides a device for detecting process parameters of a part, comprising: A part information acquisition module, the part information acquisition module is used to acquire images of the target part through an industrial camera to acquire part image information, and at the same time acquire part laser point cloud information through a three-dimensional scanner; a part space model acquisition module, the part space model acquisition module is used to spatially align and fuse the part image information and the part laser point cloud information to generate a target part space model, and at the same time obtain a standard part space model based on the drawing information of the target part; a multi-dimensional loss analysis module, the multi-dimensional loss analysis module is used to perform multi-dimensional loss analysis on the target part space model and the standard part space model to obtain part surface deviation parameters; a part process detection feature parameter acquisition module, the part process detection feature parameter acquisition module is used to acquire internal defect feature parameters of the target part through ultrasonic sensor detection, and acquire part process detection feature parameters based on the part surface deviation parameters and the internal defect feature parameters; a part production optimization control module, the part production optimization control module is used to perform process parameter compensation optimization based on the part process detection feature parameters to obtain part production process optimization parameters, and perform production optimization control on the target part through the part production process optimization parameters.

[0006] In a possible implementation, the target part space model is generated by performing the following processing: The industrial camera and the three-dimensional scanner are calibrated separately to determine the camera calibration parameters and the scanner calibration parameters; joint calibration is performed based on the camera calibration parameters and the scanner calibration parameters to establish an image pixel-laser point cloud mapping relationship; key feature points are extracted from the part image information and the part laser point cloud information in turn to obtain an image feature point set and a point cloud feature point set; feature point matching is performed on the image feature point set and the point cloud feature point set according to the image pixel-laser point cloud mapping relationship to obtain part feature point matching information; data fusion and three-dimensional reconstruction are performed on the part image information and the part laser point cloud information based on the part feature point matching information to generate the target part space model.

[0007] In a possible implementation, the part surface deviation parameter is obtained by performing the following processing: According to the part process requirement standard, a feature twin network is constructed, and the feature twin network includes a standard twin network and a target twin network, and the standard twin network and the target twin network are shared weight networks; based on the standard twin network and the target twin network, multi-dimensional feature extraction is performed on the standard part space model and the target part space model respectively, and a standard part feature set and a target part feature set are output; a feature loss function is introduced to perform loss analysis on the standard part feature set and the target part feature set to obtain the part surface deviation parameters.

[0008] In a possible implementation, the part production process optimization parameters are obtained and the following processing is performed: A part process feature simulation model is established, and process control compensation analysis is performed on the part process detection feature parameters based on the part process feature simulation model to obtain a process control parameter selection threshold; a plurality of process control parameters are randomly generated based on the process control parameter selection threshold, and simulation evaluation is performed on the plurality of process control parameters through the part process feature simulation model to obtain a plurality of part process feature prediction qualities; clustering, expansion and optimization are performed on the plurality of process control parameters according to the plurality of part process feature prediction qualities to obtain the part production process optimization parameters.

[0009] In a possible implementation, the part process feature simulation model is established by performing the following processing: The production process parameter space of the target part is constructed by data mining technology, and the production process parameter space includes production process control parameters and corresponding part process detection feature data; the part process detection feature data is classified into multi-dimensional indicators according to the part process requirement standard to obtain a part process feature indicator classification data set; the production process control parameters are respectively associated and fitted with the part process feature indicator classification data set to obtain a process control parameter-process indicator feature model set; each associated model in the process control parameter-process indicator feature model set is subjected to equal-weight simulation fusion to establish the part process feature simulation model.

[0010] In a possible implementation, the part production process optimization parameters are obtained by performing the following processing: The multiple process control parameters are optimized according to the predicted qualities of the multiple part process features to obtain multiple parent process control parameters with preset proportions; the multiple parent process control parameters are respectively used as cluster centers to cluster the multiple process control parameters to obtain multiple process control parameter clusters; based on the multiple parent process control parameters, the multiple process control parameter clusters are expanded and optimized to obtain multiple process control optimization parameter clusters; based on the multiple process control optimization parameter clusters, global comparison and optimization are performed to output the part production process optimization parameters.

[0011] In a possible implementation, the obtaining of multiple process control optimization parameter clusters involves performing the following processing: Based on the multiple parent process control parameters, preset intra-cluster distance parameters are screened within the multiple process control parameter clusters to obtain multiple intra-cluster adjacent control parameters; the multiple parent process control parameters and the multiple intra-cluster adjacent control parameters are cross-combined and parameter mutated to obtain multiple child process control parameters; based on the multiple child process control parameters, the multiple process control parameter clusters are expanded with child parameters to obtain multiple process control expanded parameter clusters; iterative optimization and update are performed within the multiple process control expanded parameter clusters to obtain the multiple process control optimized parameter clusters.

[0012] The present application also provides a method for detecting process parameters of a part, comprising: The target part is imaged by an industrial camera to obtain part image information, and the part laser point cloud information is obtained by a three-dimensional scanner; the part image information and the part laser point cloud information are spatially aligned and fused to generate a target part space model, and a standard part space model is obtained based on the drawing information of the target part; a multi-dimensional loss analysis is performed on the target part space model and the standard part space model to obtain part surface deviation parameters; the internal defect characteristic parameters of the target part are obtained by ultrasonic sensor detection, and the part process detection characteristic parameters are obtained based on the part surface deviation parameters and the internal defect characteristic parameters; process parameter compensation optimization is performed based on the part process detection characteristic parameters to obtain part production process optimization parameters, and the target part is subjected to production optimization control through the part production process optimization parameters.

[0013] The invention proposes a device and method for detecting process parameters of a part proposed in the present application. The industrial camera in the part information acquisition module is used to acquire the image of the target part to acquire the part image information. The laser point cloud information of the part is acquired by a three-dimensional scanner at the same time. The part image information and the laser point cloud information of the part are spatially registered and fused by the part space model acquisition module to generate a target part space model. At the same time, the standard part space model is obtained according to the drawing information of the target part. The multi-dimensional loss analysis module is used to perform multi-dimensional loss analysis on the target part space model and the standard part space model to obtain the part surface deviation parameters. The ultrasonic sensor in the part process detection feature parameter acquisition module is used to detect and acquire the internal defect feature parameters of the target part. The part process detection feature parameters are acquired according to the part surface deviation parameters and the internal defect feature parameters. The part production optimization control module is used to perform process parameter compensation optimization based on the part process detection feature parameters to obtain the part production process optimization parameters. The target part is then subjected to production optimization control by the part production process optimization parameters, thereby achieving the technical effect of efficient and accurate detection of parts, adaptive optimization of production processes, and improving product process quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the device according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0015] Figure 1 A schematic diagram of the structure of a device for detecting process parameters of a part provided in an embodiment of the present application.

[0016] Figure 2 A schematic flow chart of a method for detecting process parameters of a part provided in an embodiment of the present application.

[0017] Explanation of the reference numerals: part information acquisition module 10 , part space model acquisition module 20 , multi-dimensional loss analysis module 30 , part process detection characteristic parameter acquisition module 40 , part production optimization control module 50 . DETAILED DESCRIPTION

[0018] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0019] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0020] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions. For example, a process, method, device, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0021] The present application embodiment provides a device for detecting process parameters of a part, such as Figure 1 As shown, the device comprises: Part information acquisition module 10, the part information acquisition module 10 is used to capture images of target parts through industrial cameras to obtain part image information, and to simultaneously obtain part laser point cloud information through three-dimensional scanners. Specifically, an industrial camera (a high-precision, high-speed image acquisition device used to obtain two-dimensional image information of an object) is used to photograph the target part to obtain two-dimensional image information of the part. At the same time, a three-dimensional scanner (a device that can obtain three-dimensional coordinate information on the surface of an object, achieved through laser technology) is used to scan the target part, and the three-dimensional coordinate information of the part surface is obtained by receiving the reflected laser signal to form laser point cloud data, which contains geometric information such as the shape and size of the part surface.

[0022] Part space model acquisition module 20, the part space model acquisition module 20 is used to perform spatial registration and fusion of the part image information and the part laser point cloud information to generate a target part space model, and at the same time obtain a standard part space model based on the drawing information of the target part. Specifically, the collected image and point cloud data are pre-processed by denoising, filtering, etc. to improve the accuracy and reliability of the data. The pre-processed image and point cloud data are aligned and fused through a spatial registration algorithm to form a target part space model in a unified spatial coordinate system. According to the drawing information of the target part, a space model of the part under standard conditions is constructed using CAD software or other three-dimensional modeling tools.

[0023] In a possible implementation, the generating of the target part space model includes: calibrating the industrial camera and the three-dimensional scanner respectively to determine the camera calibration parameters and the scanner calibration parameters; performing joint calibration based on the camera calibration parameters and the scanner calibration parameters to establish an image pixel-laser point cloud mapping relationship; extracting key feature points of the part image information and the part laser point cloud information in turn to obtain an image feature point set and a point cloud feature point set; performing feature point matching on the image feature point set and the point cloud feature point set according to the image pixel-laser point cloud mapping relationship to obtain part feature point matching information; performing data fusion and three-dimensional reconstruction on the part image information and the part laser point cloud information based on the part feature point matching information to generate the target part space model.

[0024] Specifically, a specific calibration plate or calibration object is used, and multiple images of the calibration plate at different angles or positions are taken. The internal parameters (such as focal length, optical center position, etc.) and external parameters (such as the position and posture of the camera relative to the world coordinate system) of the industrial camera are calculated using image processing algorithms. These parameters are the camera calibration parameters. For a 3D scanner, the calibration parameters of the scanner are adjusted by scanning an object of known shape and size and comparing the scan result with the actual size of the object. These parameters are the scanner calibration parameters. The industrial camera and the 3D scanner are installed together or fixed in relative position, and the mapping relationship between the image pixels and the laser point cloud is established by simultaneously photographing and scanning the same object, and using the known camera and scanner calibration parameters. That is, through joint calibration, it is determined which position in the point cloud each pixel in the image corresponds to.

[0025] The image and point cloud data are processed by image processing algorithms and point cloud processing algorithms, and points with significant features (such as corners, edges, etc.) are extracted as feature points. Using the established image pixel-laser point cloud mapping relationship, the image feature point set is matched with the point cloud feature point set to find the corresponding relationship between them. The matched image feature points and point cloud feature points are fused to form a unified data representation. Based on the fused data, the 3D reconstruction algorithm is used to generate the spatial model of the target part through steps such as point cloud splicing and surface reconstruction. This implementation method ensures the accuracy of the measurement data through the calibration and joint calibration of the industrial camera and 3D scanner, thereby improving the accuracy of 3D reconstruction.

[0026] The multidimensional loss analysis module 30 is used to perform multidimensional loss analysis on the target part space model and the standard part space model to obtain part surface deviation parameters. Specifically, the target part space model is compared and analyzed with the standard part space model, including comparisons in terms of shape, size, position, etc. A loss function or error measurement method is introduced to calculate the difference between the target part and the standard part to obtain part surface deviation parameters, which include deviation value, deviation direction, deviation distribution, etc.

[0027] In a possible implementation, the obtaining of part surface deviation parameters includes: building a feature twin network according to part process requirement standards, the feature twin network including a standard twin network and a target twin network, and the standard twin network and the target twin network are shared weight networks; based on the standard twin network and the target twin network, performing multi-dimensional feature extraction on the standard part space model and the target part space model respectively, and outputting a standard part feature set and a target part feature set; introducing a feature loss function to perform loss analysis on the standard part feature set and the target part feature set to obtain the part surface deviation parameters.

[0028] Specifically, in order to compare the difference between the standard part space model and the target part space model, a feature twin network is built according to the process parameter standards such as the part's geometric shape, dimensional accuracy, flatness, surface defects, etc. The network consists of two sub-networks with the same structure and shared weights, namely the standard twin network and the target twin network. The standard part space model and the target part space model are input into the standard twin network and the target twin network respectively. Through the network layers such as the convolution layer and the pooling layer, the spatial features (shape, size, surface quality, etc.) of the parts are extracted. Each twin network outputs a feature set, which represents the multi-dimensional features of the standard part and the target part respectively. Using the defined feature loss function, the loss value between the standard part feature set and the target part feature set is calculated. Through the back propagation algorithm, the network weight is adjusted to minimize the feature loss. When the feature loss reaches the preset threshold or the number of training times reaches the preset upper limit, the training is stopped and the part surface deviation parameters are output. This implementation method extracts the high-dimensional features of the parts by building a feature twin network, and measures the difference between the standard part and the target part through the loss function, so as to achieve accurate detection of part surface deviation.

[0029] Part process detection characteristic parameter acquisition module 40, the part process detection characteristic parameter acquisition module 40 is used to obtain the internal defect characteristic parameters of the target part through ultrasonic sensor detection, and obtain the part process detection characteristic parameters according to the part surface deviation parameters and the internal defect characteristic parameters. Specifically, the target part is internally inspected by an ultrasonic sensor, and the internal defects of the part are detected by emitting ultrasonic waves and receiving reflected signals, such as cracks, pores, inclusions and other structures or materials that do not meet the requirements. According to the detection results of the ultrasonic sensor, the internal defect characteristic parameters of the part are extracted, such as the defect position, size, shape, etc. The part surface deviation parameters and the internal defect characteristic parameters are integrated to form a complete set of part process detection characteristic parameters.

[0030] The part production optimization control module 50 is used to perform process parameter compensation optimization based on the part process detection characteristic parameters, obtain part production process optimization parameters, and perform production optimization control on the target part through the part production process optimization parameters. Specifically, based on the part process detection characteristic parameters, the manufacturing process parameters (such as temperature, pressure, time, etc.) are compensated and optimized to reduce the deviation and defects of the parts during the manufacturing process. Based on the compensated and optimized process parameters, the part production process optimization parameters are generated. The production process optimization parameters are applied to the production process of the target parts, and production optimization control is achieved by adjusting parameters such as machinery and equipment and process conditions. The embodiment of the present application utilizes an industrial camera and a three-dimensional scanner to simultaneously collect image information and laser point cloud information of a part, spatially aligns and fuses the image information and laser point cloud information of the part, generates a target part space model, and performs multi-dimensional loss analysis with the standard part space model to obtain part surface deviation parameters, obtains internal defect characteristic parameters of the part through ultrasonic sensor detection, obtains part process detection characteristic parameters in combination with the part surface deviation parameters, performs process parameter compensation optimization based on the part process detection characteristic parameters, obtains part production process optimization parameters, and uses these parameters to perform production optimization control on the target part, thereby achieving the technical effect of efficient and accurate detection of parts, adaptive optimization of production processes, and improving product process quality.

[0031] In one possible implementation, obtaining the part production process optimization parameters includes: establishing a part process feature simulation model, performing process control compensation analysis on the part process detection feature parameters based on the part process feature simulation model, and obtaining a process control parameter selection threshold; randomly generating multiple process control parameters based on the process control parameter selection threshold, and performing simulation evaluation on the multiple process control parameters through the part process feature simulation model to obtain multiple part process feature prediction qualities; clustering, expanding and optimizing the multiple process control parameters according to the multiple part process feature prediction qualities to obtain the part production process optimization parameters.

[0032] Specifically, in order to accurately simulate the part production process and its results, a simulation model that can reflect the relationship between the part process characteristics and production conditions is established. By collecting historical production data, process parameters, part quality data, etc., according to the characteristics of the part production process, simulation tools are used to design the simulation model structure, input the collected data into the model, set the initial parameters of the model, and verify the accuracy and reliability of the model by comparing with the actual production data. Considering the process control parameter deviation caused by factors such as equipment errors, compensation analysis is performed on the part process detection characteristic parameters, that is, the influence of factors such as equipment errors and process fluctuations on the process control parameters is analyzed. According to the analysis results, the compensation strategy is determined, such as adjusting the range of process control parameters or introducing correction factors. Based on the compensation strategy, the selection threshold of the process control parameters is set, that is, within the range of process control parameters, after considering factors such as equipment errors, a reasonable parameter value range or limit is determined to ensure that the compensated process control parameters can meet the part process requirements.

[0033] According to the set threshold, the value range of the process control parameters is determined. Within the value range, a random number generator is used to generate multiple combinations of process control parameters. The generated process control parameter combinations are input into the part process feature simulation model, the simulation model is run, the production process of the part is simulated, and the simulation results, including the predicted quality of the process features of the part, are collected. The simulation results are clustered using a clustering algorithm, similar process control parameter combinations are classified into one category, and the average process feature prediction quality of each type of process control parameter combination is calculated. According to the average quality, the optimal or better process control parameter combination is selected, and further search and optimization are performed near the optimal parameters to obtain more accurate optimization parameters as the part production process optimization parameters. This implementation method achieves accurate evaluation and optimization of process control parameters by establishing a part process feature simulation model and simulating the production process and results of the part.

[0034] In one possible implementation, the establishment of a part process feature simulation model includes: constructing a production process parameter space of the target part through data mining technology, wherein the production process parameter space includes production process control parameters and corresponding part process detection feature data; performing multi-dimensional index classification on the part process detection feature data according to the part process requirement standard to obtain a part process feature index classification data set; correlating and fitting the production process control parameters with the part process feature index classification data set respectively to obtain a process control parameter-process index feature model set; performing equal-weighted simulation fusion on each associated model in the process control parameter-process index feature model set to establish the part process feature simulation model.

[0035] Specifically, historical data related to the production of target parts is collected from production records, quality inspection reports and other sources. The collected data is preprocessed by cleaning, deduplication, normalization and other preprocessing operations to ensure the quality and consistency of the data. The preprocessed data is processed using data mining techniques (such as cluster analysis, association rule mining, etc.) to construct a production process parameter space containing production process control parameters (such as temperature, pressure, time, etc.) and corresponding part process detection feature data (such as size, hardness, surface roughness, etc.). According to the specific requirements set for various process indicators (such as dimensional accuracy, material properties, etc.) in the part production process, the specific standards and dimensions of multi-dimensional (i.e., multiple process detection features) indicator classification are determined, and the part process detection feature data is classified using classification algorithms (such as decision trees, support vector machines, etc.) to obtain a classification data set of part process feature indicators. By using the correlation fitting method (such as linear regression, nonlinear regression, neural network, etc.), the production process control parameters are correlated with the part process characteristic index classification data set to train the process control parameter-process indicator characteristic model set, which is a set of models, each of which describes the correlation between a specific process control parameter and the corresponding process indicator characteristic. Each correlation model in the process control parameter-process indicator characteristic model set is fused with equal weights to obtain a comprehensive simulation model, namely the part process characteristic simulation model, which can predict the part process characteristics according to the production process control parameters. This implementation method constructs the production process parameter space through data mining technology, fully utilizes the information in the historical data, and fully utilizes the advantages of each model by integrating the prediction results of multiple models, reducing the deviation and error that may exist in a single model, and improving the overall prediction accuracy.

[0036] In a possible implementation, obtaining the part production process optimization parameters includes: optimizing the multiple process control parameters according to the predicted qualities of the multiple part process features to obtain multiple parent process control parameters with preset proportions; clustering the multiple process control parameters using the multiple parent process control parameters as cluster centers to obtain multiple process control parameter clusters; expanding and optimizing the multiple process control parameter clusters based on the multiple parent process control parameters to obtain multiple process control optimization parameter clusters; performing global comparison and optimization based on the multiple process control optimization parameter clusters to output the part production process optimization parameters.

[0037] Specifically, according to the prediction quality of multiple part process features (these qualities are obtained by simulating and evaluating the randomly generated process control parameters through the part process feature simulation model), multiple process control parameters are optimized, and those process control parameters that can produce higher prediction quality are selected, for example, the top 10% or 20% excellent parameters are selected as parents. These optimized parent process control parameters are respectively used as cluster centers, and the remaining process control parameters are clustered to group similar parameters together to form multiple process control parameter clusters. The parameters in each cluster are similar to some extent. In each formed process control parameter cluster, the parent parameters of the cluster center are expanded and optimized, including adding new parameters to the cluster, fine-tuning the cluster center parameters, etc., to explore the possible better parameter combinations in the cluster. After completing the expansion and optimization of all clusters, all the obtained process control optimization parameter clusters are globally compared to find the parameter combination with the best prediction quality in all clusters. This combination is determined as the part production process optimization parameter. This implementation method quickly narrows the search scope through optimization and clustering operations, reduces the number of process control parameter combinations that need to be evaluated, and improves the efficiency of the optimization process. By performing expansion optimization within each cluster, it fully utilizes the similarity of parameters within the cluster. At the same time, global comparison ensures that the final selected parameter combination is the best among all possible options, thereby enhancing the optimization effect.

[0038] In one possible implementation, the obtaining of multiple process control optimization parameter clusters includes: performing preset intra-cluster distance parameter screening within the multiple process control parameter clusters based on the multiple parent process control parameters to obtain multiple intra-cluster adjacent control parameters; performing cross-combination and parameter variation on the multiple parent process control parameters and the multiple intra-cluster adjacent control parameters to obtain multiple child process control parameters; performing child parameter expansion on the multiple process control parameter clusters based on the multiple child process control parameters to obtain multiple process control expansion parameter clusters; and performing iterative optimization and updating within the multiple process control expansion parameter clusters to obtain the multiple process control optimization parameter clusters.

[0039] Specifically, according to multiple parent process control parameters, in each process control parameter cluster, the distance between each parent parameter and other parameters is calculated (this distance can be Euclidean distance, Manhattan distance, etc., depending on the dimension and nature of the parameter). According to the preset distance threshold, the parameters that are close to each parent parameter are screened out, and these screened out parameters are the adjacent control parameters within the cluster. Randomly select the parent parameter and the adjacent parameter within the cluster, exchange part of their genes (i.e., part of the parameter) to generate new offspring parameters. The offspring parameters generated after the crossover combination are adjusted randomly or based on probability to introduce new mutations and increase the diversity of solutions. The generated offspring parameters are added to their corresponding process control parameter clusters, so that each cluster contains more parameters, forming a process control expansion parameter cluster. Repeat the above steps (starting from the preset intra-cluster distance parameter screening), but each iteration is based on the current optimal or approximately optimal process control parameter cluster. In each iteration, the process control parameter cluster is updated according to the new offspring parameters, and the global comparison and optimization are re-performed. The iterative process continues until the stopping condition is reached (such as the number of iterations, performance improvement is less than a certain threshold, etc.). Through iterative optimization and updating, the quality of the process control parameter cluster is continuously improved until the optimal or approximately optimal process control optimization parameter cluster is found. This implementation adopts the idea of ​​genetic algorithm, and through steps such as crossover combination, parameter variation and iterative optimization, it enhances the global search capability, increases the diversity of solutions, helps to find better solutions in the search process, and avoids premature convergence.

[0040] In the above, refer to Figure 1 A process parameter detection device for a part according to an embodiment of the present invention is described in detail. Figure 2 A method for detecting process parameters of a part according to an embodiment of the present invention is described.

[0041] A method for detecting process parameters of a part according to an embodiment of the present invention is used to solve the technical problems of insufficient detection accuracy and low detection efficiency in existing part process parameter detection, so as to achieve the technical effect of efficiently and accurately detecting parts, realizing adaptability optimization of production processes, and improving product process quality.

[0042] A method for detecting process parameters of a part comprises: acquiring image information of a target part by an industrial camera, and acquiring laser point cloud information of the part by a three-dimensional scanner; spatially registering and fusing the part image information and the part laser point cloud information to generate a target part space model, and acquiring a standard part space model according to the drawing information of the target part; performing multi-dimensional loss analysis on the target part space model and the standard part space model to obtain part surface deviation parameters; acquiring internal defect characteristic parameters of the target part by ultrasonic sensor detection, and acquiring part process detection characteristic parameters according to the part surface deviation parameters and the internal defect characteristic parameters; performing process parameter compensation optimization based on the part process detection characteristic parameters to obtain part production process optimization parameters, and performing production optimization control on the target part by using the part production process optimization parameters.

[0043] Among them, the generating of the target part space model may further include: calibrating the industrial camera and the three-dimensional scanner respectively to determine the camera calibration parameters and the scanner calibration parameters; performing joint calibration based on the camera calibration parameters and the scanner calibration parameters to establish an image pixel-laser point cloud mapping relationship; extracting key feature points of the part image information and the part laser point cloud information in turn to obtain an image feature point set and a point cloud feature point set; performing feature point matching on the image feature point set and the point cloud feature point set according to the image pixel-laser point cloud mapping relationship to obtain part feature point matching information; performing data fusion and three-dimensional reconstruction on the part image information and the part laser point cloud information based on the part feature point matching information to generate the target part space model.

[0044] Among them, the obtaining of part surface deviation parameters may further include: building a feature twin network according to the part process requirement standard, the feature twin network including a standard twin network and a target twin network, and the standard twin network and the target twin network are shared weight networks; based on the standard twin network and the target twin network, performing multi-dimensional feature extraction on the standard part space model and the target part space model respectively, and outputting a standard part feature set and a target part feature set; introducing a feature loss function to perform loss analysis on the standard part feature set and the target part feature set to obtain the part surface deviation parameters.

[0045] Among them, the obtaining of part production process optimization parameters may further include: establishing a part process feature simulation model, performing process control compensation analysis on the part process detection feature parameters based on the part process feature simulation model, and obtaining a process control parameter selection threshold; randomly generating multiple process control parameters based on the process control parameter selection threshold, and performing simulation evaluation on the multiple process control parameters through the part process feature simulation model to obtain multiple part process feature prediction qualities; clustering, expanding and optimizing the multiple process control parameters according to the multiple part process feature prediction qualities to obtain the part production process optimization parameters.

[0046] Among them, the establishment of a part process feature simulation model may further include: constructing a production process parameter space of the target part through data mining technology, the production process parameter space including production process control parameters and corresponding part process detection feature data; performing multi-dimensional index classification on the part process detection feature data according to the part process requirement standard to obtain a part process feature index classification data set; correlating and fitting the production process control parameters with the part process feature index classification data set respectively to obtain a process control parameter-process index feature model set; performing equal-weight simulation fusion on each associated model in the process control parameter-process index feature model set to establish the part process feature simulation model.

[0047] Among them, obtaining the part production process optimization parameters may further include: optimizing the multiple process control parameters according to the predicted qualities of the multiple part process characteristics to obtain multiple parent process control parameters with preset proportions; clustering the multiple process control parameters using the multiple parent process control parameters as cluster centers to obtain multiple process control parameter clusters; expanding and optimizing the multiple process control parameter clusters based on the multiple parent process control parameters to obtain multiple process control optimization parameter clusters; performing global comparison and optimization based on the multiple process control optimization parameter clusters to output the part production process optimization parameters.

[0048] Wherein, obtaining a plurality of process control optimization parameter clusters may further include: Based on the multiple parent process control parameters, preset intra-cluster distance parameters are screened within the multiple process control parameter clusters to obtain multiple intra-cluster adjacent control parameters; the multiple parent process control parameters and the multiple intra-cluster adjacent control parameters are cross-combined and parameter mutated to obtain multiple child process control parameters; based on the multiple child process control parameters, the multiple process control parameter clusters are expanded with child parameters to obtain multiple process control expanded parameter clusters; iterative optimization and update are performed within the multiple process control expanded parameter clusters to obtain the multiple process control optimized parameter clusters.

[0049] A device for detecting process parameters of a part provided in an embodiment of the present invention can execute a method for detecting process parameters of a part provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0050] Although the present application makes various references to certain modules in the device according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0051] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A device for detecting process parameters of a part, characterized in that: The device comprises: A part information acquisition module, which is used to acquire images of target parts through an industrial camera to acquire part image information, and to acquire part laser point cloud information through a three-dimensional scanner; A part space model acquisition module, which is used to perform spatial registration and fusion of the part image information and the part laser point cloud information to generate a target part space model, and at the same time obtain a standard part space model according to the drawing information of the target part; A multidimensional loss analysis module, which is used to perform multidimensional loss analysis on the target part space model and the standard part space model to obtain part surface deviation parameters; A part process detection characteristic parameter acquisition module, the part process detection characteristic parameter acquisition module is used to acquire the internal defect characteristic parameters of the target part through ultrasonic sensor detection, and acquire the part process detection characteristic parameters according to the part surface deviation parameters and the internal defect characteristic parameters; A part production optimization control module, wherein the part production optimization control module is used to perform process parameter compensation optimization based on the part process detection characteristic parameters, obtain part production process optimization parameters, and perform production optimization control on the target part through the part production process optimization parameters.

2. A device for detecting process parameters of a part according to claim 1, characterized in that: The generating of the target part space model comprises: Calibrate the industrial camera and the three-dimensional scanner respectively to determine camera calibration parameters and scanner calibration parameters; Performing joint calibration based on the camera calibration parameters and the scanner calibration parameters to establish an image pixel-laser point cloud mapping relationship; Extract key feature points from the part image information and the part laser point cloud information in sequence to obtain an image feature point set and a point cloud feature point set; Performing feature point matching on the image feature point set and the point cloud feature point set according to the image pixel-laser point cloud mapping relationship to obtain part feature point matching information; Based on the part feature point matching information, the part image information and the part laser point cloud information are subjected to data fusion and three-dimensional reconstruction to generate the target part space model.

3. A device for detecting process parameters of a part according to claim 1, characterized in that: The obtaining of part surface deviation parameters comprises: According to the part process requirement standard, a feature twin network is constructed, wherein the feature twin network includes a standard twin sub-network and a target twin sub-network, and the standard twin sub-network and the target twin sub-network are shared weight networks; Based on the standard twin network and the target twin network, multi-dimensional feature extraction is performed on the standard part space model and the target part space model respectively, and a standard part feature set and a target part feature set are output; A feature loss function is introduced to perform loss analysis on the standard part feature set and the target part feature set to obtain the part surface deviation parameters.

4. A device for detecting process parameters of a part as claimed in claim 3, characterized in that: The obtained parts production process optimization parameters include: Establishing a part process feature simulation model, performing process control compensation analysis on the part process detection feature parameters based on the part process feature simulation model, and obtaining a process control parameter selection threshold; Randomly generate multiple process control parameters based on the process control parameter selection threshold, and simulate and evaluate the multiple process control parameters through the part process feature simulation model to obtain multiple part process feature prediction qualities; The plurality of process control parameters are clustered, expanded and optimized according to the predicted qualities of the plurality of part process features to obtain the part production process optimization parameters.

5. A device for detecting process parameters of a part as claimed in claim 4, characterized in that: The step of establishing a part process feature simulation model comprises: Constructing the production process parameter space of the target part by data mining technology, wherein the production process parameter space includes production process control parameters and corresponding part process detection feature data; Perform multi-dimensional index classification on the part process detection feature data according to the part process requirement standard to obtain a part process feature index classification data set; Correlate and fit the production process control parameters with the part process characteristic index classification data set to obtain a process control parameter-process index characteristic model set; Each associated model in the process control parameter-process indicator feature model set is subjected to equal-weighted simulation fusion to establish the part process feature simulation model.

6. A device for detecting process parameters of a part according to claim 4, characterized in that: The step of obtaining the optimized parameters of the parts production process includes: Optimizing the plurality of process control parameters according to the predicted qualities of the plurality of part process features to obtain a plurality of parent process control parameters of a preset proportion; Taking the multiple parent process control parameters as cluster centers respectively, clustering the multiple process control parameters to obtain multiple process control parameter clusters; Based on the multiple parent process control parameters, expansion optimization is performed on the multiple process control parameter clusters to obtain multiple process control optimization parameter clusters; A global comparison and optimization is performed based on the multiple process control optimization parameter clusters to output the part production process optimization parameters.

7. A device for detecting process parameters of a part according to claim 6, characterized in that: The obtaining of multiple process control optimization parameter clusters includes: Based on the multiple parent process control parameters, a preset intra-cluster distance parameter is screened within the multiple process control parameter clusters to obtain multiple intra-cluster neighboring control parameters; Cross-combining and parameter mutating the plurality of parent-generation process control parameters and the plurality of adjacent control parameters within the cluster to obtain a plurality of child-generation process control parameters; Expanding the plurality of process control parameter clusters with child parameters based on the plurality of child process control parameters to obtain a plurality of process control expansion parameter clusters; Iterative optimization and updating are performed within the plurality of process control expansion parameter clusters to obtain the plurality of process control optimization parameter clusters.

8. A method for detecting process parameters of a part, characterized in that: The method is implemented by a process parameter detection device for a part according to any one of claims 1 to 7, and the method comprises: The target parts are imaged by industrial cameras to obtain part image information, and the laser point cloud information of the parts is obtained by 3D scanners. Performing spatial registration and fusion of the part image information and the part laser point cloud information to generate a target part space model, and obtaining a standard part space model according to the drawing information of the target part; Performing multi-dimensional loss analysis on the target part space model and the standard part space model to obtain part surface deviation parameters; Acquire the internal defect characteristic parameters of the target part through ultrasonic sensor detection, and acquire the part process detection characteristic parameters according to the part surface deviation parameters and the internal defect characteristic parameters; Based on the part process detection characteristic parameters, process parameter compensation optimization is performed to obtain part production process optimization parameters, and production optimization control of the target part is performed through the part production process optimization parameters.

Images