Intelligent detection device and detection method for anti-corrosion aluminum alloy part

Abstract

The invention relates to the technical field of intelligent detection, and discloses an intelligent detection device and detection method for an anti-corrosion aluminum alloy part, and the method comprises the steps: carrying out the feature extraction of an image sequence in an anti-corrosion aluminum alloy cutting process, and obtaining a multi-dimensional feature map; performing three-dimensional space mapping and morphological feature analysis on the multi-dimensional feature map to obtain spatio-temporal feature data of the cutting trajectory; inputting the spatio-temporal feature data of the cutting trajectory into a layered cutting parameter prediction model for layered feature parameter prediction to obtain a cutting parameter prediction result; performing Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction result to obtain probability density distribution data of a cutting trajectory, and performing variable observation domain optimization on a Gaussian process model to obtain an optimization sequence of cutting parameters; and the optimized sequence of the cutting parameters is input into a Kalman filtering-Hungary algorithm for state tracking, a control instruction of the cutting parameters is obtained, then real-time accurate control over the cutting parameters is achieved, and the cutting machining quality is remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent detection technology, and in particular to an intelligent detection device and detection method for corrosion-resistant aluminum alloy parts. Background Art

[0002] In the cutting process of corrosion-resistant aluminum alloy, due to the complexity of material properties and processing environment, traditional cutting parameter control methods are difficult to meet high-precision processing requirements.

[0003] At present, the quality inspection of the anti-corrosion aluminum alloy cutting process mainly relies on manual experience and fixed parameter control. This method is not only inefficient, but also difficult to cope with dynamic changes in the cutting process. Especially under complex cutting conditions, real-time tracking of the cutting trajectory and parameter optimization are difficult, which can easily lead to problems such as unstable cutting quality and precision fluctuations. Summary of the invention

[0004] The present application provides an intelligent detection device and detection method for corrosion-resistant aluminum alloy parts, thereby realizing real-time and precise control of cutting parameters and significantly improving the cutting process quality.

[0005] The first aspect of the present application provides an intelligent detection method for corrosion-resistant aluminum alloy parts, and the intelligent detection method for corrosion-resistant aluminum alloy parts comprises: Feature extraction is performed on the image sequence of the anti-corrosion aluminum alloy cutting process to obtain a multi-dimensional feature map; Performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature graph to obtain spatiotemporal feature data of the cutting trajectory; Inputting the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; Performing Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain probability density distribution data of the cutting trajectory, and performing variable observation domain optimization on the Gaussian process model to obtain an optimized sequence of cutting parameters; The optimized sequence of cutting parameters is input into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters.

[0006] The second aspect of the present application provides an intelligent detection device for corrosion-resistant aluminum alloy parts, and the intelligent detection device for corrosion-resistant aluminum alloy parts comprises: A feature extraction module is used to extract features from the image sequence of the corrosion-resistant aluminum alloy cutting process to obtain a multi-dimensional feature map; A spatial mapping module, used for performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature map to obtain spatiotemporal feature data of the cutting trajectory; A parameter prediction module, used for inputting the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; An optimization module is used to perform Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain probability density distribution data of the cutting trajectory, and to perform variable observation domain optimization on the Gaussian process model to obtain an optimized sequence of cutting parameters; The state tracking module is used to input the optimized sequence of cutting parameters into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters.

[0007] Compared with the prior art, the present application has the following beneficial effects: through the design of a multi-scale cascade GhostNet network and a spatial positioning and feature generalization enhancement module, the feature extraction capability of complex scenes in the cutting process is significantly enhanced, and the cutting trajectory detection accuracy is greatly improved; a hierarchical cutting parameter prediction model is used for hierarchical feature fusion, combined with the design of a multi-level prediction head, to achieve accurate prediction of cutting parameters and effectively reduce the error of parameter prediction; Gaussian mixture modeling and kernel density analysis methods are introduced, and the performance degradation problem of traditional methods in low-speed cutting scenarios is solved through a variable observation domain adaptive optimization strategy, thereby ensuring the stability of cutting parameter optimization; a state tracking framework based on the Kalman filter-Hungarian algorithm is designed, combined with a parameter optimization controller and a trajectory feedback compensation mechanism, to achieve real-time and precise control of cutting parameters and significantly improve the cutting processing quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0009] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0010] Figure 1 It is a schematic flow chart of an intelligent detection method for corrosion-resistant aluminum alloy parts provided by an embodiment of the present invention; Figure 2It is a schematic block diagram of the structure of an intelligent detection device for corrosion-resistant aluminum alloy parts provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0012] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0013] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0014] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. Figure 1 , an embodiment of the intelligent detection method of the corrosion-resistant aluminum alloy parts in the embodiment of the present application includes: Step 100, extracting features from an image sequence of the corrosion-resistant aluminum alloy cutting process to obtain a multi-dimensional feature map; It is understandable that the execution subject of the present application may be an intelligent detection device for corrosion-resistant aluminum alloy parts, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0015] Specifically, the image sequence of the anti-corrosion aluminum alloy cutting process is subjected to multi-scale pyramid decomposition to obtain high-resolution feature layer, medium-resolution feature layer and low-resolution feature layer. Multi-scale pyramid decomposition is to decompose the input image into image levels of different resolutions through downsampling and convolution operations. Each level contains feature representations of different scales from the input image, where the high-resolution feature layer retains more detail information, the medium-resolution feature layer focuses on the salient pattern of the local area, and the low-resolution feature layer extracts global low-frequency semantic information. The high-resolution feature layer is input into the first-level GhostNet module for feature extraction. The design of the GhostNet module is based on the idea of ​​efficient feature representation, and the Ghost bottleneck unit is used to combine the trunk processing of information with the decomposition of redundant features. The first-level GhostNet module contains 8 Ghost bottleneck units, each of which consists of a main branch convolution layer and a parallel branch transformation layer. The main branch convolution layer extracts basic features, and the parallel branch transformation layer generates redundant features through simple linear operations, thereby reducing the amount of calculation while retaining the detail information of the image. Through this step, the first-level feature map is extracted from the high-resolution feature layer to capture the micro-texture and edge features in the cutting process. The medium-resolution feature layer is input into the second-level GhostNet module for feature extraction. The second-level GhostNet module contains 6 Ghost bottleneck units. In order to improve the accuracy of feature extraction and the expressiveness of the model, a channel attention mechanism and a spatial attention mechanism are embedded in each bottleneck unit. The channel attention mechanism enables the model to pay more attention to the key features related to the cutting process and ignore redundant information by assigning weights to different channels; the spatial attention mechanism enhances the model's ability to capture salient areas in the image by weighting the spatial position. After being processed by this module, the generated second-level feature map can effectively express the salient patterns of local areas during the cutting process. The low-resolution feature layer is input into the third-level GhostNet module for feature extraction. The third-level GhostNet module consists of 4 Ghost bottleneck units. In order to improve the modeling ability of global semantic information, each bottleneck unit is configured with cross-layer residual connections and adaptive feature aggregation units. Cross-layer residual connections can effectively alleviate the gradient vanishing problem that occurs during information transmission, while ensuring the fusion between low-level features and high-level features. The adaptive feature aggregation unit enhances the global modeling ability of low-resolution features by dynamically adjusting the weights of features of different scales. The third-level feature map generated by the third-level GhostNet module mainly expresses the macro information and global trends in the cutting process. The first-level feature map, the second-level feature map, and the third-level feature map are fused. A weighted fusion strategy is used in the fusion process to make the information of features with different resolutions complementary and generate a fused feature map containing multi-scale information. The fused feature map is input into the feature enhancement module for processing, which contains 3 depth-separable convolutional layers and 1 global context encoder.The depthwise separable convolution layer greatly reduces the computational complexity by decomposing the standard convolution into depthwise convolution and pointwise convolution, while retaining the image detail information. Each depthwise separable convolution layer is followed by a batch normalization layer to standardize the feature distribution and introduce nonlinearity through the PReLU activation function to improve the feature representation capability. The global context encoder enhances the model's perception of global information by modeling the global contextual relationship of the entire image. After being processed by the feature enhancement module, a multi-dimensional feature map is obtained.

[0016] Step 200, performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature graph to obtain spatiotemporal feature data of the cutting trajectory; Specifically, the multidimensional feature map is transformed into spatial coordinates. The two-dimensional image features are converted into position information in a three-dimensional space coordinate system. Through the spatial transformation method based on projective geometry, combined with the actual calibration parameters of the cutting equipment and the geometric relationship of the multi-view images, the spatial mapping matrix of the cutting trajectory is generated. The matrix takes the pixel position of the cutting trajectory as input and maps its actual position in the three-dimensional coordinate system, thereby accurately representing the dynamic changes of the cutting trajectory in the three-dimensional space. According to the spatial mapping matrix, the geometric feature extraction operator of the cutting trajectory is constructed to analyze the geometric features of the cutting trajectory. The geometric feature extraction operator is the core tool of feature analysis, including the directional gradient operator, the curvature operator and the continuity operator. Among them, the directional gradient operator is used to calculate the rate of change of the cutting trajectory in different directions, thereby extracting the edge characteristics of the trajectory; the curvature operator reflects the curvature of the cutting trajectory based on the local curvature change of the trajectory points; the continuity operator ensures the complete description of the cutting trajectory by evaluating the connectivity between the trajectory points. Through the comprehensive analysis of these geometric operators, the geometric feature vector of the cutting trajectory is generated to represent the overall shape of the cutting trajectory and its local change characteristics. Morphological operations are performed on the geometric feature vectors to extract the morphological feature data of the cutting trajectory. Morphological operations are an image processing method based on set theory, and operations include expansion, erosion, opening operations, and closing operations. In this process, by performing structured operations on the geometric feature vectors, the key features of the trajectory are enhanced, redundant information is removed, and the morphological boundaries of the cutting trajectory are made clearer. Temporal feature analysis is performed on the morphological feature data. Based on the time series distribution of the trajectory points, the time series feature sequence of the cutting trajectory is generated by analyzing the dynamic change pattern of the trajectory points on the time axis. This sequence is based on time and quantitatively describes the law of the change of the geometric shape of the trajectory over time. On this basis, a spatiotemporal feature association model is established based on the time series feature sequence of the cutting trajectory. This model completes the spatiotemporal association modeling of the cutting trajectory by combining the spatial position of the trajectory with the time evolution process. A method based on a graph neural network or a recurrent neural network is used to jointly model the spatial features and time series features of the cutting trajectory, and extract the association features of the trajectory in the two dimensions of time and space. Through spatiotemporal association modeling, the dynamic behavior pattern and spatial distribution law of the cutting trajectory are captured, and the spatiotemporal association features of the cutting trajectory are generated. The spatial mapping matrix, morphological feature data and spatiotemporal correlation features of the cutting trajectory are combined to generate the spatiotemporal feature data of the cutting trajectory.

[0017] Step 300: input the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; It should be noted that a hierarchical cutting parameter prediction model is constructed, and the input spatiotemporal feature data is decomposed layer by layer through the feature classification network to extract feature information at different levels. The spatiotemporal feature data of the cutting trajectory is input into the feature classification network, and the network decomposes the input feature data into local feature layer, regional feature layer and global feature layer through a series of convolution operations and pooling processes. The local feature layer is used to capture the detailed information of the cutting trajectory, the regional feature layer focuses on the characteristic analysis of the local area, and the global feature layer covers the overall spatial and temporal dynamics of the cutting trajectory. After completing the feature classification, the local feature layer is input into the first prediction head in the hierarchical cutting parameter prediction model for processing. The first prediction head contains 4 convolution blocks and 1 fully connected layer, each of which is equipped with a spatial pyramid pooling module. The spatial pyramid pooling module effectively captures the multi-scale information of local features by performing pooling operations on features at multiple scales, thereby improving the ability to characterize the details of the cutting trajectory. The stacking of convolution blocks further extracts features in depth, and the fully connected layer maps the extracted features into local cutting parameter prediction data, which are used to describe the microscopic changes in the tool trajectory during the cutting process and the local parameter adjustment requirements. At the same time, the regional feature layer is input into the second prediction head for processing. The second prediction head contains 3 convolution blocks and 1 fully connected layer, and each convolution block is configured with a channel cascade module. The channel cascade module enhances the model's ability to capture the dependencies between channels by establishing a cascade relationship between different channels, so that the regional feature layer can better express the significant patterns of local areas in the cutting trajectory. The layer-by-layer stacking of convolution blocks further refines the features to ensure the accurate expression of regional information, and the fully connected layer maps the extracted regional features into regional cutting parameter prediction data. These prediction data are used to guide the optimization of parameters in the mesoscale range during the cutting process, such as cutting speed and energy distribution. At the same time, the global feature layer is input into the third prediction head for processing. The third prediction head consists of 2 convolution blocks and 1 fully connected layer, and each convolution block is set with a feature recalibration module. The feature recalibration module dynamically adjusts the global features by learning the importance weights of the features, so that the model pays more attention to the key information related to the overall state of the cutting trajectory. The introduction of this module can significantly improve the sensitivity of the model to global information and ensure the accurate modeling of the overall morphology and spatiotemporal dynamics of the cutting trajectory. The fully connected layer maps the global features to global cutting parameter prediction data, which are used to describe the adjustment of the overall strategy during the cutting process, such as the global optimization of the cutting path and the balanced distribution of parameters. The local cutting parameter prediction data, the regional cutting parameter prediction data and the global cutting parameter prediction data are input into the parameter fusion network for fusion processing. The parameter fusion network integrates the detail accuracy of local features, the regional information of regional features and the integrity of global features by combining data at different levels.During the fusion process, strategies such as weighted averaging, attention mechanism or multi-layer perceptron are used to dynamically assign weights according to the importance of features at different levels to generate cutting parameter prediction results.

[0018] Step 400: Perform Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain the probability density distribution data of the cutting trajectory, and perform variable observation domain optimization on the Gaussian process model to obtain the optimized sequence of cutting parameters; Specifically, a Gaussian mixture model is constructed based on the prediction results of cutting parameters. By analyzing the spatial characteristics of the cutting trajectory, the trajectory is divided into several local areas, and a Gaussian component is generated for each local area. Each Gaussian component contains a mean vector and a covariance matrix, which together define the position and distribution shape of the component in the cutting trajectory. By constructing a parameter group containing N Gaussian components, the multimodal distribution characteristics of the cutting trajectory are preliminarily described. The expectation maximization algorithm is performed according to the Gaussian component parameter group to optimize the model parameters. The expectation maximization algorithm contains two main steps. In the "expectation" step, for each cutting trajectory data point, the posterior probability of belonging to each Gaussian component is calculated, and these probability values ​​indicate the attribution weight of the data point on the components in different regions. In the "maximization" step, the parameters of the Gaussian components, including the mean vector and the covariance matrix, are re-estimated using the calculated posterior probability, and these parameters are optimized through continuous iteration. This optimization process continuously improves the log-likelihood function value of the model in each iteration until convergence. The optimized Gaussian mixture parameters are obtained to accurately describe the probability distribution of the cutting trajectory. The Gaussian mixture parameters are input into the radial basis kernel function for density calculation. The radial basis kernel function can smooth the probability distribution and improve the accuracy of density estimation. The bandwidth parameter matrix of the kernel function is set, which determines the smoothness and local sensitivity of the kernel function. In order to select the optimal bandwidth parameter, the cross-validation optimization method is adopted. By grid searching the likelihood function values ​​under different bandwidth parameters, the bandwidth parameter that maximizes the likelihood function value is selected to obtain the target kernel density function. After obtaining the target kernel density function, probability compensation is performed for the special needs of the low-speed cutting area. The low-speed cutting area is identified by the target kernel density function, which is manifested as a low-value area in the density distribution. The compensation weights are calculated based on these areas. The weight calculation is based on the local physical characteristics of the trajectory and the cutting stability requirements. The original probability density value is updated using these compensation weights to generate the compensated density function. In order to ensure the overall consistency and physical meaning of the compensated density function, the probability integral normalization is performed on it, that is, the density value is divided by the overall density integral, so that the density distribution maintains the normalized characteristics in the global range. The probability density distribution data of the cutting trajectory is obtained to describe the probability characteristics of the cutting trajectory in space. Based on the generated probability density distribution data, the Gaussian process model is optimized with variable observation domain. In this process, the input space of the Gaussian process model is dynamically adjusted to adapt to the distribution characteristics of different regions in the cutting trajectory. In the low-speed cutting area, a compensation density function is introduced to adjust the weight distribution of the observation domain, thereby enhancing the prediction ability of the model in these areas. At the same time, by optimizing the covariance function of the model, the adaptability and robustness of the Gaussian process model to the uncertainty area are improved. The optimized Gaussian process model can more accurately predict the dynamic changes of the cutting parameters. The optimized sequence of cutting parameters is generated by the above-mentioned optimized Gaussian process model.

[0019] The probability density distribution data of the cutting trajectory is input into the observation domain partitioning module for adaptive partitioning processing. By analyzing the spatial probability distribution characteristics of the cutting trajectory, the cutting trajectory is dynamically partitioned according to the density changes and characteristic requirements of different regions, and variable observation domain configuration data is generated. This process uses density thresholds and regional clustering algorithms to divide the cutting trajectory into several sub-regions, each of which has independent observation domain characteristics, ensuring that the subsequent optimization process can be accurately adjusted according to the characteristics of the local region. The kernel function of the Gaussian process model is constructed according to the variable observation domain configuration data. The design of the kernel function adopts a combined kernel function based on Markov properties. The kernel function captures the characteristics of the periodic change pattern and the random change pattern in the cutting trajectory by combining the weighted combination of the periodic kernel function and the Gaussian kernel function. The periodic kernel function ensures the accurate prediction of the repeated cutting pattern by modeling the periodic behavior of the trajectory, while the Gaussian kernel function uses its smoothness characteristics and local correlation to capture random disturbances and regional differences. In the process of constructing the kernel function, the weight parameters are dynamically adjusted so that the combined kernel function can adapt the generated kernel function parameters according to the characteristics of different regions of the cutting trajectory. The kernel function parameters of the Gaussian process model are optimized for hyperparameters. The core of hyperparameter optimization is to ensure the modeling accuracy of the Gaussian process model for the cutting trajectory characteristics by searching for the optimal parameter combination. The optimization process combines the physical constraints of the cutting process to search for parameters, including the cutting speed, the smoothness of the tool path, and the restrictions on material properties. Through methods such as grid search or Bayesian optimization, the key hyperparameters of the kernel function, such as the amplitude factor, the correlation scale, and the weight coefficient, are gradually adjusted, and the model is evaluated based on the logarithmic marginal likelihood value to obtain the optimized Gaussian process model. Based on the optimized Gaussian process model, an uncertainty-based active learning strategy is used to select sampling points. The active learning strategy identifies the key points that need further sampling by analyzing the prediction uncertainty of the model for different cutting parameter spaces. At these key points, sparse sampling is performed to maximize the information gain and generate a target sampling point sequence. Based on the target sampling point sequence, the cutting quality is evaluated and the sampling points are scored to generate a cutting parameter scoring matrix. The cutting parameter scoring matrix combines the parameter accuracy predicted by the model, the local quality characteristics of the sampling points, and the actual effect of the cutting process, and quantitatively evaluates different sampling points with a multi-dimensional scoring mechanism. The scoring matrix is ​​input into the sequence optimizer to perform time-series optimization of the cutting parameters. The sequence optimizer selects the optimal parameter path based on the scoring matrix through a dynamic programming algorithm or a reinforcement learning method to achieve global optimization of the cutting parameters. During the selection process, the optimizer fully considers the continuity and globality of the cutting parameters to ensure that the final optimization sequence has a smooth change trend at different time points while meeting the operational constraints in the actual cutting process. The cutting parameter optimization sequence is obtained through the above steps.

[0020] Step 500: Input the optimized sequence of cutting parameters into the Kalman filter-Hungarian algorithm for state tracking to obtain control instructions for the cutting parameters.

[0021] Specifically, the optimized sequence of cutting parameters is input into the state prediction module to construct a state space model for cutting parameter tracking. The core of the state prediction module is to establish the system state equation and the measurement equation, which respectively describe the state transfer process and observation process of the cutting parameters. By constructing the state transfer matrix and the observation matrix of the cutting parameters, the dynamic evolution of the cutting parameters in the time series dimension and the relationship between the measured value and the actual state are characterized to form a state space model of the cutting parameters. The state space model of the cutting parameters is calculated by Kalman filtering. The first step of Kalman filtering is the prediction stage, which calculates the predicted state vector of the cutting parameters through the state transfer matrix and the control input, and estimates the predicted covariance matrix to measure the trust of the system in the predicted state. After obtaining the predicted state vector, it is combined with the actual observation data to correct the system state. The optimized sequence and the observation data are jointly input into the Hungarian algorithm module, which uses the bipartite graph optimal matching strategy to associate multiple target states. The Hungarian algorithm generates the target association matrix by optimizing the cost matrix between the predicted state vector and the observation data, and clarifies the best matching relationship between each observation value and the predicted state. The association result is measured and updated. The measurement update process includes calculating the Kalman gain matrix, the posterior state estimate, and the posterior covariance matrix. The Kalman gain matrix is ​​an important parameter for weighing the predicted state and the actual observed value. Its calculation is based on the ratio of the predicted covariance and the observed noise covariance, and is used to dynamically adjust the weight of the state correction. After the Kalman gain matrix is ​​calculated, it is applied to the target association matrix to correct the predicted state, and the posterior state estimate, that is, the optimal state sequence of the cutting parameters, is obtained. At the same time, the posterior covariance matrix is ​​updated to reflect the uncertainty of the system. The optimal state sequence of the cutting parameters is input into the parameter optimization controller to achieve the joint optimization of the key parameters of the cutting process. The parameter optimization controller analyzes the cutting speed, feed rate, and cutting depth in the optimal state sequence, and uses a multi-objective optimization algorithm to comprehensively adjust these parameters to generate preliminary parameter adjustment instructions. These instructions are designed to ensure that the cutting process achieves a balance between efficiency and quality while meeting material properties and process requirements. Trajectory feedback compensation is performed on the parameter adjustment instructions. The process of trajectory feedback compensation is based on the error calculation between the real-time cutting state and the target trajectory. By comparing the current state with the target state, a compensation value is calculated, which can correct the deviation caused by system error or external interference. After the compensation value is calculated, it is superimposed on the parameter adjustment instruction to generate the final cutting parameter control instruction.

[0022] In the embodiments of the present application, through the design of a multi-scale cascade GhostNet network and a spatial positioning and feature generalization enhancement module, the feature extraction capability of complex scenes in the cutting process is significantly enhanced, and the cutting trajectory detection accuracy is greatly improved; a hierarchical cutting parameter prediction model is used for hierarchical feature fusion, combined with the design of a multi-level prediction head, to achieve accurate prediction of cutting parameters and effectively reduce the error of parameter prediction; Gaussian mixture modeling and kernel density analysis methods are introduced, and the performance degradation problem of traditional methods in low-speed cutting scenarios is solved through a variable observation domain adaptive optimization strategy, thereby ensuring the stability of cutting parameter optimization; a state tracking framework based on the Kalman filter-Hungarian algorithm is designed, combined with a parameter optimization controller and a trajectory feedback compensation mechanism, to achieve real-time and precise control of cutting parameters and significantly improve the cutting process quality.

[0023] In a specific embodiment, the process of executing step 100 may specifically include the following steps: Multi-scale pyramid decomposition is performed on the image sequence of the anti-corrosion aluminum alloy cutting process to obtain high-resolution feature layer, medium-resolution feature layer and low-resolution feature layer; The high-resolution feature layer is input into the first-level GhostNet module for feature extraction. The first-level GhostNet module contains 8 Ghost bottleneck units, each of which consists of a main branch convolution layer and a parallel branch transformation layer to obtain the first-level feature map; The medium-resolution feature layer is input into the second-level GhostNet module for feature extraction. The second-level GhostNet module contains 6 Ghost bottleneck units. Each Ghost bottleneck unit sets a channel attention mechanism and a spatial attention mechanism to obtain a second-level feature map. The low-resolution feature layer is input into the third-level GhostNet module for feature extraction. The third-level GhostNet module contains 4 Ghost bottleneck units. Each Ghost bottleneck unit is configured with cross-layer residual connection and adaptive feature aggregation unit to obtain the third-level feature map. The first-level feature map, the second-level feature map and the third-level feature map are fused to obtain a fused feature map, and the fused feature map is input into the feature enhancement module for processing. The feature enhancement module contains 3 depth-wise separable convolutional layers and 1 global context encoder. Each depth-wise separable convolutional layer is followed by a batch normalization layer and a PReLU activation function to obtain a multidimensional feature map.

[0024] Specifically, the image sequence of the cutting process is used as input data for multi-scale pyramid decomposition. The original image data is decomposed into high-resolution feature layer, medium-resolution feature layer and low-resolution feature layer through downsampling and Gaussian filtering, which correspond to the image details, local patterns and global structure respectively. Assume that the input image sequence is ,in and represents the spatial coordinates of the image, Represents the time frame. The multi-scale decomposition is expressed by the formula: ; ; ; in is the high-resolution feature layer, is the medium-resolution feature layer, is a low-resolution feature layer, DownSample() is a downsampling operation, which is implemented using convolution or mean pooling with a stride of 2. Input the first level GhostNet module for feature extraction. The GhostNet module reduces the computational cost through the Ghost bottleneck unit while maintaining the feature expression capability. Each Ghost bottleneck unit consists of a main branch convolution layer and a parallel branch transformation layer. The main branch convolution layer is used to generate basic features, denoted as , the parallel branch generates redundant features through simple linear transformation, denoted as The feature extraction process is expressed as: ; ; ; in and is the convolution kernel and transformation weight matrix, is the bias term, and They are nonlinear activation function and linear transformation function respectively, and * indicates convolution operation. After being processed by 8 Ghost bottleneck units, a high-resolution feature map is obtained. Similarly, the medium-resolution feature layer Enter the second-level GhostNet module. The second-level GhostNet module contains 6 Ghost bottleneck units, and introduces a channel attention mechanism and a spatial attention mechanism in each unit. The channel attention mechanism learns the importance weight of each feature channel. , to enhance key channels: ; Where GAP() is the global average pooling operation, and are the weights and biases of the fully connected layer. The spatial attention mechanism generates spatial weights To enhance the features of salient areas: ; The final feature fusion is: ; Get the medium resolution feature map For low-resolution feature layers , and input it into the third-level GhostNet module, which contains 4 Ghost bottleneck units. Each unit is configured with cross-layer residual connections and adaptive feature aggregation units. Cross-layer residual connections achieve direct flow of information by short-circuiting input and output, and are defined as: ; The adaptive feature aggregation unit achieves fusion by dynamically adjusting the weights of features with different resolutions: ; in and is the adaptive weight parameter, and Pooling() is the downsampling operation. After processing, a low-resolution feature map is obtained. After completing the feature extraction of the three resolutions, the first level feature map , the second level feature map And the third-level feature map Perform feature fusion. Feature fusion is achieved through weighted accumulation: ; in is the fusion weight, which is used to adjust the contribution ratio of features with different resolutions. The input feature enhancement module is processed. The feature enhancement module consists of 3 depth-wise separable convolutional layers and 1 global context encoder. The depth-wise separable convolutional layer reduces the computational complexity by decomposing the convolution operation. Its operation is defined as: ; ; in and The global context encoder enhances the features by modeling the global feature relationship and finally generates a multi-dimensional feature map. .

[0025] In a specific embodiment, the process of executing step 200 may specifically include the following steps: Perform spatial coordinate transformation on the multi-dimensional feature map, convert the two-dimensional image features into position information in a three-dimensional spatial coordinate system, and obtain the spatial mapping matrix of the cutting trajectory; According to the space mapping matrix, a geometric feature extraction operator of the cutting trajectory is constructed. The geometric feature extraction operator includes a directional gradient operator, a curvature operator and a continuity operator. The geometric feature analysis of the cutting trajectory is performed to obtain a geometric feature vector of the cutting trajectory. Performing morphological operations on the geometric feature vectors to obtain morphological feature data of the cutting trajectory, and performing time series feature analysis on the morphological feature data to obtain a time series feature sequence of the cutting trajectory; A spatiotemporal feature correlation model is established based on the temporal feature sequence of the cutting trajectory, and the spatial position and time evolution of the cutting trajectory are modeled to obtain the spatiotemporal correlation features of the cutting trajectory; The spatial mapping matrix of the cutting trajectory, the morphological feature data of the cutting trajectory and the spatiotemporal correlation features of the cutting trajectory are combined to obtain the spatiotemporal feature data of the cutting trajectory.

[0026] Specifically, the spatial coordinate transformation is performed on the multi-dimensional feature map. Assume that the input two-dimensional image feature is ,in and is the spatial coordinate of the image. The image coordinates are mapped to the three-dimensional space coordinate system through the camera's intrinsic matrix and extrinsic matrix. The intrinsic matrix Describes the focal length and principal point offset of the camera, the external parameter matrix Including rotation matrix and translation vectors The mapping relationship is expressed as: ; in is a point in three-dimensional space, are the homogeneous coordinates of the points in the image, is the internal parameter matrix, is the external parameter matrix. Through the above transformation, each two-dimensional image feature point Map the position information into three-dimensional space to generate the spatial mapping matrix of the cutting trajectory , which is of the form: ; in is the total number of feature points. After obtaining the spatial mapping matrix, the geometric feature extraction operator of the cutting trajectory is constructed to perform geometric feature analysis on the trajectory. The geometric feature extraction operator includes the directional gradient operator, the curvature operator, and the continuity operator. Among them, the directional gradient operator is used to calculate the local direction change of each point on the trajectory, which is defined as: ; in and represents the spatial variation between adjacent points, Yes The curvature operator is used to calculate the local curvature of a point on the trajectory, and its formula is: ; in Yes The curvature of represents the curvature of its trajectory. The continuity operator is used to evaluate the continuity of trajectory points and is defined as: ; in Yes The Euclidean distance to the next point reflects the smoothness of the trajectory. By applying the above operator to all points on the trajectory, the geometric feature vector of the cutting trajectory is obtained. , describing the direction change, curvature and coherence characteristics of the trajectory. After completing the geometric feature analysis, morphological operations are performed on the geometric feature vector. Morphological operations include operations such as expansion, erosion and skeletonization to enhance the morphological characteristics of the trajectory. For example, the local edge of the trajectory is expanded by the expansion operation, the noise points are removed by the erosion operation, and the skeletonization is used to extract the skeleton path of the trajectory. Let the structural element of the morphological operation be The trajectory characteristics are , the expansion operation is expressed as: ; The corrosion operation is: ; After morphological operation, the morphological feature data of the cutting trajectory is obtained. The temporal feature analysis is performed on the morphological feature data to extract the temporal dynamic change pattern of the trajectory. The temporal feature analysis is completed by constructing the temporal difference vector: ; in and The time and The morphological characteristics of the moment, Is the corresponding time series feature vector. Time series feature vector based on cutting trajectory , establish a spatiotemporal feature association model, and extract the spatiotemporal association features of the trajectory by modeling the spatial position and time evolution of the trajectory. The spatiotemporal association model is represented by a spatiotemporal graph, where nodes are trajectory points and edges are spatiotemporal association weights between points. The spatiotemporal association weight is defined as: ; in and It is a trajectory point and spatial location, It is a parameter that controls the range of association. The spatial mapping matrix of the cutting trajectory , morphological feature data and spatiotemporal correlation features Combine to generate spatiotemporal feature data of cutting trajectory For example, during the cutting process, if irregular high-frequency vibration occurs in a certain area, the abnormal area is found through curvature and timing characteristics, and the abnormal point is associated with its neighborhood through the spatiotemporal correlation model, and finally a global feature matrix reflecting the dynamic changes of the trajectory is generated.

[0027] In a specific embodiment, the process of executing step 300 may specifically include the following steps: The spatiotemporal feature data of the cutting trajectory are input into the feature classification network in the hierarchical cutting parameter prediction model for feature classification to obtain a local feature layer, a regional feature layer and a global feature layer; The local feature layer is input into the first prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain local cutting parameter prediction data. The first prediction head includes 4 convolution blocks and 1 fully connected layer. A spatial pyramid pooling module is set in each convolution block. The regional feature layer is input into the second prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain regional cutting parameter prediction data. The second prediction head includes 3 convolution blocks and 1 fully connected layer, and each convolution block is configured with a channel cascade module. The global feature layer is input into the third prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain global cutting parameter prediction data. The third prediction head includes 2 convolution blocks and 1 fully connected layer. Each convolution block is provided with a feature recalibration module. The local cutting parameter prediction data, the regional cutting parameter prediction data and the global cutting parameter prediction data are input into the parameter fusion network for parameter fusion processing to obtain the cutting parameter prediction results.

[0028] Specifically, the input cutting trajectory spatiotemporal feature data The data is passed to the feature classification network for hierarchical processing. The core of the feature classification network is to decompose the spatiotemporal feature data into three levels: local, regional and global through a series of convolution operations and pooling operations to obtain the local feature layer. , Regional feature layer and the global feature layer The feature classification process is expressed by the following formula: ; ; ; in, It is a convolution operation, used to extract local features; and They are the pooling operations of regional and global features, respectively, using maximum pooling or average pooling to reduce the resolution and expand the receptive field. After completing the feature classification, the local feature layer The first prediction head in the layered cutting parameter prediction model is input for processing. The first prediction head contains 4 convolution blocks and 1 fully connected layer. Each convolution block integrates a spatial pyramid pooling (SPP) module to extract multi-scale information. convolutional blocks, whose output is expressed as: ; ; in, Represents pooling operations at different scales, and concat[] represents concatenation of features at different scales. , output local cutting parameter prediction data: ; At the same time, the regional feature layer Input the second prediction head, which contains 3 convolution blocks and 1 fully connected layer. Each convolution block is configured with a channel cascade module. The channel cascade module enhances the feature expression capability by building dependencies between channels. The formula is: ; ; in, is an activation function (such as Sigmoid), is the global average pooling operation, and are weights and biases respectively. The regional features are passed through the fully connected layer Output area cutting parameter prediction data: ; Global feature layer It is passed to the third prediction head, which contains 2 convolution blocks and 1 fully connected layer. Each convolution block is configured with a feature recalibration module. The feature recalibration module dynamically weights the features by adjusting their importance. The process is as follows: ; in, and is the weight parameter obtained through learning, is a pooling operation. The global features are passed through the fully connected layer Output global cutting parameter prediction data: ; After completing the above steps, the local, regional and global cutting parameter prediction data are , and Input parameter fusion network for fusion processing. The parameter fusion network adopts a multi-layer perceptron structure and achieves the final prediction by weighted accumulation of features in each layer. The fusion process is: ; in, It is the fusion weight obtained through optimization training. For example, assuming that the speed and depth of the tool need to be optimized during the cutting process, the local feature reflects the local overheating of a certain area, the regional feature captures the heat distribution relationship between the area and the adjacent area, and the global feature provides the trend information of the overall heat diffusion.

[0029] In a specific embodiment, the process of executing step 400 may specifically include the following steps: Based on the cutting parameter prediction results, N Gaussian components are constructed, and the N Gaussian components include a mean vector and a covariance matrix, wherein each Gaussian component corresponds to a local area of ​​the cutting trajectory, and a Gaussian component parameter group is obtained; Performing an expectation maximization iteration operation according to the Gaussian component parameter group, the expectation maximization iteration operation includes calculating the posterior probability that each data point belongs to each Gaussian component and updating the parameters of the Gaussian components to obtain optimized Gaussian mixture parameters; The optimized Gaussian mixture parameters are input into the radial basis kernel function for density calculation, and the kernel function bandwidth parameter matrix is ​​set to obtain the initial kernel density function. The initial kernel density function is cross-validated and optimized. The likelihood function values ​​under different bandwidth parameters are grid searched, and the bandwidth parameter corresponding to the maximized likelihood function value is selected to obtain the target kernel density function. Based on the target kernel density function, the probability compensation of the low-speed cutting area is performed. The probability compensation process includes identifying the low-speed cutting area, calculating the compensation weight and updating the probability density value to obtain the compensation density function, and performing probability integral normalization processing on the compensation density function. The probability density value is divided by the overall density integral to obtain the probability density distribution data of the cutting trajectory. The Gaussian process model is optimized in variable observation domain according to the probability density distribution data of the cutting trajectory, and the optimized sequence of cutting parameters is obtained.

[0030] Specifically, based on the prediction results of the cutting trajectory, the trajectory is locally partitioned and constructed Gaussian components, each of which is used to describe the probability distribution characteristics of a local area in the cutting trajectory. Assume that the cutting trajectory data is ,in Indicates the trajectory data points -dimensional features. The probability density function of the Gaussian component is expressed as: ; in, It is The mean vector of the Gaussian components, is the covariance matrix, Represents its determinant, is a parameter group. By local partitioning of the cutting trajectory, Gaussian components are combined to form a Gaussian mixture model (GMM). The overall probability density of GMM is: ; in It is The weights of the Gaussian components satisfy , is the parameter set of all Gaussian components. After constructing the Gaussian components, their parameters need to be optimized. The expectation maximization (EM) algorithm is introduced to iteratively update Maximize the log-likelihood function: ; The EM algorithm is divided into two stages: the expectation (E) step and the maximization (M) step. In the E step, the Belong to The posterior probability (responsibility) of the Gaussian components is: ; in Represents data points In the The probability weights on the Gaussian components. In the M step, according to Update model parameters: ; ; By repeatedly iterating the above steps until the model converges, the optimized Gaussian mixture parameters are obtained. The optimized Gaussian mixture parameters are input into the radial basis kernel function (RBF) for density calculation. The form of RBF is: ; in is the bandwidth parameter of the kernel function, which is used to control the smoothness of the density estimate. The initial kernel density function is obtained by a weighted combination of the kernel function and the Gaussian component: ; To optimize the bandwidth parameters , cross-validate the initial kernel density function for different bandwidth values, and select the bandwidth parameter that maximizes the log-likelihood function value through grid search The optimized density function is: ; Based on the target kernel density function, probability compensation is performed on the low-speed cutting area. The low-speed cutting area passes the density threshold. Identify, when satisfy When , the point is judged as a low-speed area. In these areas, the compensation weight is calculated: ; And update the probability density value: ; The compensated density function is then integrated and normalized to ensure that the density sums to 1: ; The probability density distribution data of the cutting trajectory is used for the variable observation domain optimization of the Gaussian process model. By adjusting the observation domain range of the Gaussian process, the kernel function and covariance function of the model are optimized according to the density value of the local area, and the optimization sequence of cutting parameters is generated.

[0031] In a specific embodiment, the execution step performs variable observation domain optimization on the Gaussian process model according to the probability density distribution data of the cutting trajectory to obtain the optimized sequence of cutting parameters may specifically include the following steps: The probability density distribution data of the cutting trajectory is input into the observation domain partitioning module to perform adaptive partitioning of the cutting trajectory, and obtain variable observation domain configuration data; The kernel function of the Gaussian process model is constructed according to the variable observation domain configuration data. The kernel function is a combined kernel function based on the Markov property, including a weighted combination of a periodic kernel function and a Gaussian kernel function, and the kernel function parameters of the Gaussian process model are obtained. Hyperparameter optimization is performed on the kernel function parameters of the Gaussian process model, and parameter search is performed based on the physical constraints of the cutting process to obtain the optimized Gaussian process model; Based on the optimized Gaussian process model, an active learning strategy based on uncertainty is used to select sampling points, and sparse sampling is performed on the cutting parameter space to obtain the target sampling point sequence; The cutting quality is evaluated and the sampling points are scored based on the target sampling point sequence to obtain a cutting parameter scoring matrix, which is then input into a sequence optimizer for cutting parameter timing optimization, and the optimal parameter path is selected to obtain an optimized sequence of cutting parameters.

[0032] Specifically, the probability density distribution data of the cutting trajectory is input into the observation domain partitioning module to perform adaptive partitioning of the cutting trajectory. The probability density distribution data of the cutting trajectory is expressed as ,in is a spatial point in the trajectory. By setting the density threshold , the cutting track is divided into high-density area and low-density area. Assume that the high-density area and low-density areas satisfy: ; In order to refine the observation domain division, each area is divided into sub-areas based on clustering algorithms (such as K-Means or DBSCAN), and the variable observation domain configuration data is generated in combination with the temporal characteristics of the trajectory. , used to describe the scope and characteristics of each sub-area. After completing the observation domain division, configure the data according to the variable observation domain Constructs the kernel function of the Gaussian process model. Kernel function It is the core of the Gaussian process and defines the similarity between data points. In order to adapt to the dynamic changes of the cutting trajectory, the kernel function adopts a combination of Markov properties, including periodic kernel function and Gaussian kernel function A weighted combination of: ; in and is a weight parameter that satisfies . Periodic kernel function The periodic pattern used to capture the cutting trajectory is of the form: ; in is the amplitude parameter, is the length scale, is a period. Gaussian kernel function It is used to model the smoothness of local areas in the trajectory: ; According to the variable observation domain configuration data, the kernel function parameters are optimized for each sub-region separately to generate a regional adaptive kernel function parameter group. After obtaining the kernel function of the Gaussian process model, the kernel function parameters are hyperparameter optimized. The optimization goal is to maximize the logarithmic marginal likelihood function of the model: ; in is the observed value, is the input data, is the covariance matrix generated by the kernel function, is a hyperparameter of the kernel function. The optimal parameter combination is searched by gradient descent or Bayesian optimization, and the physical constraints of the cutting process (such as the limit of cutting speed and depth) are combined to ensure the physical rationality of the model and obtain an optimized Gaussian process model. Based on the optimized Gaussian process model, an active learning strategy based on uncertainty is used to select sampling points. The prediction uncertainty of the Gaussian process model is determined by the prediction mean and the prediction variance represents, where: ; ; By selecting points with high prediction variance for sampling, the information gain of the model can be maximized to generate the target sampling point sequence The cutting quality is evaluated based on the target sampling point sequence and a score is assigned to each sampling point. The scoring is based on cutting accuracy, surface quality and the size of the heat affected zone. The comprehensive scoring matrix It is expressed as: ; in are the scoring functions for accuracy, surface quality and thermal influence, is a weight parameter. The scoring matrix is ​​input into the sequence optimizer, which uses dynamic programming or reinforcement learning algorithms to find the optimal parameter path in the time series and generate the final cutting parameter optimization sequence. For example, during the cutting process, if the probability density of a certain area is low and the corresponding thermal impact is large, the active learning strategy will select this area for more sampling point analysis. Through the Gaussian process model to optimize and evaluate these points, the sequence optimizer finally generates a parameter adjustment strategy, such as reducing the tool speed and depth, to ensure the improvement of cutting quality and effective protection of materials.

[0033] In a specific embodiment, the process of executing step 500 may specifically include the following steps: The optimized sequence of cutting parameters is input into the state prediction module, which includes the system state equation and measurement equation, constructs the state transfer matrix and observation matrix of the cutting parameters, and obtains the state space model of the cutting parameters; The Kalman filter calculation is performed on the state space model of the cutting parameters to obtain the predicted state vector of the cutting parameters, and the predicted state vector of the cutting parameters and the actual observation data are input into the Hungarian algorithm. The Hungarian algorithm module uses the bipartite graph optimal matching strategy to associate the multi-target states and obtain the target association matrix. The target correlation matrix is ​​measured and updated, and the measurement update includes calculating the Kalman gain matrix, the posterior state estimate and the posterior covariance matrix to obtain the optimal state sequence of the cutting parameters; The optimal state sequence of cutting parameters is input into the parameter optimization controller, and the cutting speed, feed rate and cutting depth are jointly optimized to obtain parameter adjustment instructions. The parameter adjustment instructions are then subjected to trajectory feedback compensation to obtain compensation values. The compensation values ​​are superimposed on the parameter adjustment instructions to obtain control instructions for the cutting parameters.

[0034] Specifically, a state prediction module is established. The core of the state prediction module is to construct the system state equation and measurement equation to describe the dynamic changes of cutting parameters and the relationship between observation and real state. Assume that the cutting parameter optimization sequence is ,in Respectively represent cutting speed, feed rate and cutting depth in time The system state vector is expressed as , where each state component Indicates The target state is The state value at the moment. The state transition equation describes the dynamic change of the state over time, in the form of: ; in is the state transfer matrix, which represents the time dependence of the system state, is the control input matrix, which represents the influence of the optimization parameters on the state, is the process noise, is the process noise covariance matrix. The measurement equation defines the relationship between the observed value and the true state in the form of: ; in is the observation vector, is the observation matrix, which represents the mapping relationship between observation values ​​and states. is the observation noise, is the observation noise covariance matrix. By constructing the above state transfer matrix and the observation matrix , forming a state space model of the cutting parameters to describe the dynamic behavior of the system. The state space model is subjected to Kalman filtering to estimate the predicted state vector of the cutting parameters Kalman filtering includes two stages: prediction and update. In the prediction stage, the current state and covariance are predicted according to the state transfer equation: ; ; in is the predicted state vector, is the prediction covariance matrix. In the update phase, combined with the observed values Perform state correction and calculate the Kalman gain matrix: ; Then update the state estimate and covariance: ; ; in is the Kalman gain matrix, which represents the correction weight of the observation value on the state estimation. The predicted state vector and the observation data are input into the Hungarian algorithm module, and the multi-target states are associated using the bipartite graph optimal matching strategy to generate the target association matrix. Assume that the matching cost matrix is , whose value indicates the predicted state and observations The Hungarian algorithm minimizes the total cost: ; in is an element of the matching matrix (valued as 0 or 1), indicating and After generating the target correlation matrix, the measurement update is performed to obtain the optimal state sequence. The optimal state sequence is input into the parameter optimization controller to jointly optimize the cutting speed, feed rate and cutting depth. The optimization controller uses a multi-objective optimization algorithm to find the optimal solution for the control parameters. , the objective function is: ; in is the error in cutting speed, feed and depth, Is the weight. Perform trajectory feedback compensation on the parameter adjustment command and calculate the compensation value by comparing the current trajectory with the target trajectory. : ; in is the feedback gain. The compensation value is added to the parameter adjustment instruction to generate the final control instruction: ; For example, when the cutting trajectory shows abnormal speed, the Kalman filter can quickly predict its changing trend. The Hungarian algorithm will associate the actual observation value with the predicted value for correction, and adjust the speed and depth parameters in the optimization controller. Combined with the trajectory feedback compensation correction instruction, the cutting quality and stability are ensured. The generated control instructions can dynamically adapt to complex cutting conditions and realize intelligent real-time control.

[0035] The above describes the intelligent detection method of the anti-corrosion aluminum alloy parts in the embodiment of the present application. The following describes the intelligent detection device 10 for the anti-corrosion aluminum alloy parts in the embodiment of the present application. Figure 2 In the embodiment of the present application, an intelligent detection device 10 for corrosion-resistant aluminum alloy parts includes: A feature extraction module 11 is used to extract features from an image sequence of the corrosion-resistant aluminum alloy cutting process to obtain a multi-dimensional feature map; A spatial mapping module 12 is used to perform three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature map to obtain spatiotemporal feature data of the cutting trajectory; The parameter prediction module 13 is used to input the spatiotemporal characteristic data of the cutting trajectory into the hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; The optimization module 14 is used to perform Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain the probability density distribution data of the cutting trajectory, and to perform variable observation domain optimization on the Gaussian process model to obtain the optimized sequence of cutting parameters; The state tracking module 15 is used to input the optimized sequence of cutting parameters into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters.

[0036] Through the collaborative cooperation of the above components, through the design of multi-scale cascade GhostNet network and spatial positioning and feature generalization enhancement modules, the feature extraction capability of complex scenes in the cutting process is significantly enhanced, and the cutting trajectory detection accuracy is greatly improved; the hierarchical cutting parameter prediction model is used for hierarchical feature fusion, combined with the design of multi-level prediction head, to achieve accurate prediction of cutting parameters and effectively reduce the error of parameter prediction; Gaussian mixture modeling and kernel density analysis methods are introduced, and the performance degradation problem of traditional methods in low-speed cutting scenarios is solved through variable observation domain adaptive optimization strategy, ensuring the stability of cutting parameter optimization; a state tracking framework based on the Kalman filter-Hungarian algorithm is designed, combined with parameter optimization controller and trajectory feedback compensation mechanism, to achieve real-time and precise control of cutting parameters and significantly improve the cutting processing quality.

[0037] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0038] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable an electronic device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc., various media that can store program codes.

[0039] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent detection method for corrosion-resistant aluminum alloy parts, characterized in that: The method comprises: Feature extraction is performed on the image sequence of the anti-corrosion aluminum alloy cutting process to obtain a multi-dimensional feature map; Performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature graph to obtain spatiotemporal feature data of the cutting trajectory; Inputting the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; Performing Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain probability density distribution data of the cutting trajectory, and performing variable observation domain optimization on the Gaussian process model to obtain an optimized sequence of cutting parameters; The optimized sequence of cutting parameters is input into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters.

2. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 1 is characterized in that: The image sequence of the corrosion-resistant aluminum alloy cutting process is subjected to feature extraction to obtain a multi-dimensional feature map, including: Multi-scale pyramid decomposition is performed on the image sequence of the anti-corrosion aluminum alloy cutting process to obtain high-resolution feature layer, medium-resolution feature layer and low-resolution feature layer; Inputting the high-resolution feature layer into the first-level GhostNet module for feature extraction, the first-level GhostNet module comprises 8 Ghost bottleneck units, each Ghost bottleneck unit is composed of a main branch convolution layer and a parallel branch transformation layer, and obtaining a first-level feature map; Inputting the medium-resolution feature layer into the second-level GhostNet module for feature extraction, the second-level GhostNet module includes 6 Ghost bottleneck units, each Ghost bottleneck unit is set with a channel attention mechanism and a spatial attention mechanism, and a second-level feature map is obtained; Inputting the low-resolution feature layer into a third-level GhostNet module for feature extraction, wherein the third-level GhostNet module comprises four Ghost bottleneck units, each of which is configured with a cross-layer residual connection and an adaptive feature aggregation unit to obtain a third-level feature map; The first-level feature map, the second-level feature map and the third-level feature map are subjected to feature fusion to obtain a fused feature map, and the fused feature map is input into a feature enhancement module for processing. The feature enhancement module comprises three depth-wise separable convolutional layers and one global context encoder. Each depth-wise separable convolutional layer is followed by a batch normalization layer and a PReLU activation function to obtain a multidimensional feature map.

3. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 1 is characterized in that: The performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature graph to obtain spatiotemporal feature data of the cutting trajectory includes: Performing spatial coordinate transformation on the multi-dimensional feature map, converting the two-dimensional image features into position information in a three-dimensional spatial coordinate system, and obtaining a spatial mapping matrix of the cutting trajectory; According to the spatial mapping matrix, a geometric feature extraction operator of the cutting trajectory is constructed, wherein the geometric feature extraction operator includes a directional gradient operator, a curvature operator and a continuity operator, and a geometric feature analysis is performed on the cutting trajectory to obtain a geometric feature vector of the cutting trajectory; Performing morphological operations on the geometric feature vectors to obtain morphological feature data of the cutting trajectory, and performing time series feature analysis on the morphological feature data to obtain a time series feature sequence of the cutting trajectory; A spatiotemporal feature association model is established based on the temporal feature sequence of the cutting trajectory, and the spatial position and time evolution of the cutting trajectory are modeled to obtain the spatiotemporal association features of the cutting trajectory; The spatial mapping matrix of the cutting trajectory, the morphological feature data of the cutting trajectory and the spatiotemporal correlation features of the cutting trajectory are combined to obtain the spatiotemporal feature data of the cutting trajectory.

4. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 1, characterized in that: The step of inputting the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result includes: Inputting the spatiotemporal feature data of the cutting trajectory into a feature classification network in a hierarchical cutting parameter prediction model for feature classification to obtain a local feature layer, a regional feature layer and a global feature layer; Inputting the local feature layer into the first prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain local cutting parameter prediction data, wherein the first prediction head comprises 4 convolution blocks and 1 fully connected layer, and each convolution block is provided with a spatial pyramid pooling module; Inputting the regional feature layer into the second prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain regional cutting parameter prediction data, wherein the second prediction head comprises 3 convolution blocks and 1 fully connected layer, and each convolution block is configured with a channel cascade module; Inputting the global feature layer into the third prediction head in the hierarchical cutting parameter prediction model for feature processing to obtain global cutting parameter prediction data, wherein the third prediction head comprises 2 convolution blocks and 1 fully connected layer, and each convolution block is provided with a feature recalibration module; The local cutting parameter prediction data, the regional cutting parameter prediction data and the global cutting parameter prediction data are input into a parameter fusion network for parameter fusion processing to obtain a cutting parameter prediction result.

5. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 1, characterized in that: The cutting parameter prediction results are subjected to Gaussian mixture modeling and kernel density analysis to obtain probability density distribution data of the cutting trajectory, and the Gaussian process model is subjected to variable observation domain optimization to obtain an optimization sequence of cutting parameters, including: Based on the cutting parameter prediction result, N Gaussian components are constructed, wherein the N Gaussian components include a mean vector and a covariance matrix, wherein each Gaussian component corresponds to a local area of ​​the cutting trajectory, and a Gaussian component parameter group is obtained; Performing an expectation maximization iterative operation according to the Gaussian component parameter group, wherein the expectation maximization iterative operation includes calculating the posterior probability that each data point belongs to each Gaussian component and updating the parameters of the Gaussian components to obtain optimized Gaussian mixture parameters; The optimized Gaussian mixture parameters are input into the radial basis kernel function for density calculation, and the kernel function bandwidth parameter matrix is ​​set to obtain an initial kernel density function, the initial kernel density function is cross-validated and optimized, a grid search is performed on the likelihood function values ​​under different bandwidth parameters, and the bandwidth parameter corresponding to the maximized likelihood function value is selected to obtain a target kernel density function; Based on the target kernel density function, probability compensation is performed on the low-speed cutting area. The probability compensation process includes identifying the low-speed cutting area, calculating the compensation weight and updating the probability density value to obtain the compensation density function, and performing probability integral normalization processing on the compensation density function. The probability density value is divided by the overall density integral to obtain the probability density distribution data of the cutting trajectory; The Gaussian process model is optimized in a variable observation domain according to the probability density distribution data of the cutting trajectory to obtain an optimized sequence of cutting parameters.

6. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 5, characterized in that: The step of performing variable observation domain optimization on the Gaussian process model according to the probability density distribution data of the cutting trajectory to obtain an optimized sequence of cutting parameters includes: Inputting the probability density distribution data of the cutting trajectory into the observation domain partitioning module to perform adaptive partitioning of the cutting trajectory to obtain variable observation domain configuration data; Constructing a kernel function for the Gaussian process model according to the variable observation domain configuration data, wherein the kernel function is a combined kernel function based on the Markov property, including a weighted combination of a periodic kernel function and a Gaussian kernel function, to obtain kernel function parameters of the Gaussian process model; Performing hyperparameter optimization on kernel function parameters of the Gaussian process model, and performing parameter search based on physical constraints of the cutting process to obtain an optimized Gaussian process model; Based on the optimized Gaussian process model, an active learning strategy based on uncertainty is adopted to select sampling points, and sparse sampling is performed on the cutting parameter space to obtain a target sampling point sequence; Based on the target sampling point sequence, the cutting quality is evaluated and the sampling points are scored to obtain a cutting parameter scoring matrix, and the cutting parameter scoring matrix is ​​input into a sequence optimizer to perform cutting parameter timing optimization, select the optimal parameter path, and obtain an optimized sequence of cutting parameters.

7. The intelligent detection method for corrosion-resistant aluminum alloy parts according to claim 1 is characterized in that: The optimized sequence of cutting parameters is input into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters, including: Inputting the optimized sequence of cutting parameters into a state prediction module, wherein the state prediction module includes a system state equation and a measurement equation, constructing a state transfer matrix and an observation matrix of the cutting parameters, and obtaining a state space model of the cutting parameters; Performing Kalman filter calculation on the state space model of the cutting parameters to obtain a predicted state vector of the cutting parameters, and inputting the predicted state vector of the cutting parameters and actual observation data into a Hungarian algorithm. The Hungarian algorithm module uses a bipartite graph optimal matching strategy to associate multiple target states to obtain a target association matrix; Performing measurement update on the target association matrix, wherein the measurement update includes calculating a Kalman gain matrix, a posterior state estimate, and a posterior covariance matrix to obtain an optimal state sequence of cutting parameters; The optimal state sequence of the cutting parameters is input into the parameter optimization controller, the cutting speed, feed rate and cutting depth are jointly optimized to obtain parameter adjustment instructions, and trajectory feedback compensation is performed on the parameter adjustment instructions to obtain compensation values, and the compensation values ​​are superimposed on the parameter adjustment instructions to obtain control instructions for the cutting parameters.

8. An intelligent detection device for corrosion-resistant aluminum alloy parts, characterized in that: The intelligent detection method for the corrosion-resistant aluminum alloy part according to claim 1 is used to perform the intelligent detection device for the corrosion-resistant aluminum alloy part, comprising: A feature extraction module is used to extract features from the image sequence of the corrosion-resistant aluminum alloy cutting process to obtain a multi-dimensional feature map; A spatial mapping module, used for performing three-dimensional spatial mapping and morphological feature analysis on the multi-dimensional feature map to obtain spatiotemporal feature data of the cutting trajectory; A parameter prediction module, used for inputting the spatiotemporal characteristic data of the cutting trajectory into a hierarchical cutting parameter prediction model to perform hierarchical characteristic parameter prediction to obtain a cutting parameter prediction result; An optimization module is used to perform Gaussian mixture modeling and kernel density analysis on the cutting parameter prediction results to obtain probability density distribution data of the cutting trajectory, and to perform variable observation domain optimization on the Gaussian process model to obtain an optimized sequence of cutting parameters; The state tracking module is used to input the optimized sequence of cutting parameters into the Kalman filter-Hungarian algorithm for state tracking to obtain the control instructions of the cutting parameters.

Images