Production line product quality tracing method based on industrial vision

Abstract

The invention provides a production line product quality traceability method based on industrial vision, and the method comprises the steps: extracting key feature parameters, including a texture orientation angle, a texture period and a surface roughness parameter, capable of reflecting the wear degree of a tool from the preprocessed texture image data, and constructing a multi-dimensional feature vector of the surface micro texture of a product; according to the multi-dimensional feature vector of the product surface micro-texture, in combination with the corresponding relation between the tool wear degree and the surface texture feature, obtaining a wear degree estimated value of the current tool, and comparing the wear degree estimated value with a wear early warning threshold value; and for the identified suspected defective products, identifying defect types by analyzing surface textures, comprehensively analyzing material attributes and designing dimensional tolerance factors, and judging whether surface texture abnormity is caused by tool wear or product defects.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to an industrial vision production line product quality tracing method. Background Art

[0002] In the traceability of product quality on the production line, a key technical challenge is how to accurately identify subtle changes in the micro-texture of the product surface. The texture features of the product surface are often very small and difficult to observe and judge directly with the naked eye. At the same time, there is a complex correspondence between these texture changes and the degree of wear of the production tools. With the continuous use of production tools, their degree of wear will gradually increase, which will lead to corresponding changes in the surface texture of the processed products. However, this texture change may be very subtle in the early stages and difficult to be captured by traditional quality inspection methods. If these subtle changes cannot be discovered in time and their impact on product quality cannot be predicted, the best time to change tools may be missed, resulting in a decline in product quality. Therefore, how to accurately identify and quantify the changes in micro-texture on the product surface through advanced measurement and analysis technology, and establish an accurate correlation model between it and the degree of tool wear, is a key technical problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a production line product quality traceability method based on industrial vision, which mainly includes:

[0004] Use imaging equipment to perform microscopic imaging of the product surface, obtain texture image data of the machined surface under different cutting parameters, perform image preprocessing to improve image contrast and clarity, and eliminate interference factors such as material defects, clamping instability, and product defects themselves;

[0005] From the preprocessed texture image data, key characteristic parameters that can reflect the degree of tool wear are extracted, including texture orientation angle, texture period and surface roughness parameters, and a multi-dimensional feature vector of the product surface micro-texture is constructed;

[0006] Based on the historical usage data of the tool, including the number of processed products, cumulative cutting time and hardness of the processed material, the surface texture characteristics of the tool processing at different wear stages are collected and analyzed to establish the corresponding relationship between the degree of tool wear and the surface texture characteristics;

[0007] According to the multi-dimensional feature vector of the micro-texture of the product surface, combined with the correspondence between the tool wear degree and the surface texture characteristics, the estimated value of the current tool wear degree is obtained and compared with the wear warning threshold;

[0008] If the current tool wear exceeds the preset threshold, a tool replacement warning is triggered, and the batch of products is marked as suspected defective products. The wear critical point is determined by the vibration signal and cutting force change trend to determine the best time to replace the tool.

[0009] For products suspected of defects that have been identified, the defect type is identified by analyzing the surface texture, and the material properties and design dimensional tolerance factors are comprehensively analyzed to determine whether the abnormal surface texture is caused by tool wear or product defects.

[0010] Based on the judgment results, the product surface texture detection results and tool wear warning information are correlated and analyzed to build a two-way feedback loop for product quality prediction and tool life optimization. By continuously collecting production site data, the correspondence between tool wear degree and surface texture characteristics is dynamically updated and optimized.

[0011] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0012] The present invention discloses a production line product quality traceability method based on industrial vision. The method obtains product surface texture images through high-resolution imaging, extracts key characteristic parameters reflecting the degree of tool wear, and establishes a corresponding relationship between tool wear and surface texture characteristics. Combining vibration and cutting force monitoring data, the critical point of tool wear is determined, and tool replacement warning is realized. At the same time, the present invention can also distinguish between surface anomalies caused by tool wear and product defects itself, and construct a two-way feedback mechanism for product quality prediction and tool life optimization. By continuously collecting and analyzing production data and dynamically optimizing the corresponding relationship between tool wear and surface texture characteristics, a balance is achieved between product quality improvement and tool cost savings, and processing efficiency and product yield are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 The present invention is a flowchart of a method for tracing product quality of a production line based on industrial vision.

[0014] Figure 2 It is a schematic diagram of an industrial vision production line product quality traceability method of the present invention.

[0015] Figure 3 This is another schematic diagram of an industrial visual production line product quality traceability method of the present invention. DETAILED DESCRIPTION

[0016] The following will describe the technical solutions in the embodiments of the present invention in detail in conjunction with the accompanying drawings in the embodiments of the present invention. The described embodiments are only part of the embodiments of the present invention.

[0017] like Figure 1-3. In this embodiment, a method for tracing product quality of a production line using industrial vision may specifically include:

[0018] S101. Use imaging equipment to perform microscopic imaging of the product surface, obtain texture image data of the machined surface under different cutting parameters, perform image preprocessing to improve image contrast and clarity, and eliminate interference factors such as material defects, clamping instability, and product defects themselves.

[0019] An imaging device is used to perform microscopic imaging of the product surface to obtain texture image data of the product surface. The texture image data is preprocessed to adjust the pixel grayscale distribution of the texture image data. If the texture image data contains noise, the noise is removed. The surface texture features of the texture image data are highlighted according to the Canny edge detection algorithm to obtain an edge-enhanced texture image. Morphological operations are performed on the edge-enhanced texture image, including dilation and erosion, to identify and segment defective areas in the edge-enhanced texture image. Features of the defective area are extracted, including area, perimeter, roundness and grayscale statistics. The features of the defective area are input into a support vector machine classifier to determine whether the defective area is caused by material defects or clamping instability. According to the judgment result of the support vector machine classifier, texture images that are greatly affected by interference factors are screened out.

[0020] Exemplarily, high-resolution imaging equipment is used to perform microscopic imaging of the product surface, including scanning electron microscope, atomic force microscope and laser confocal microscope, and appropriate imaging equipment is selected according to the material characteristics and processing method. Imaging parameters such as the acceleration voltage, working distance and magnification of the scanning electron microscope are set to obtain texture image data of the processed surface under different cutting speeds, feed rates and cutting depth combinations. The acquired texture image data is preprocessed, and the image pixel grayscale distribution is adjusted using an adaptive histogram equalization algorithm to improve the image contrast. The salt and pepper noise in the image is removed by a median filtering algorithm to improve the image clarity. The Canny edge detection algorithm is used to highlight the surface texture features and obtain an edge-enhanced texture image. Morphological operations, including dilation and erosion, are performed on the edge-enhanced texture image to identify and segment possible defect areas. The region growing algorithm is used to further refine the boundaries of the defect area. The features of the segmented defect area are extracted, including area, perimeter, roundness and grayscale statistics. The extracted features are input into a support vector machine classifier to determine whether the defect area is a material defect or interference caused by clamping instability. According to the classification results of the support vector machine, the optimal threshold automatically calculated by the Otsu algorithm is set to screen out the texture images that are greatly affected by interference factors. These affected images are eliminated and the remaining valid texture image data is retained. The quality of the screened texture images is evaluated, and the signal-to-noise ratio and clarity index of the images are calculated to ensure the representativeness and reliability of the remaining image data. The texture image data that pass the quality evaluation is used as the data basis for the subsequent analysis of the relationship between cutting parameters and machined surface quality.

[0021] When using high-resolution imaging equipment to perform microscopic imaging of the product surface, a scanning electron microscope was selected as the imaging device, and the acceleration voltage was set to 15kV, the working distance was 12mm, and the magnification was 2000 times. The texture image data of the machined surface was obtained under the parameter combination of cutting speed 100m / min, feed rate 0.1mm / rev and cutting depth 0.5mm. The acquired texture image was preprocessed. First, the adaptive histogram equalization algorithm was applied to divide the image into 8x8 small blocks, and histogram equalization was performed in each small block. Then, the small blocks were merged by bilinear interpolation to obtain a contrast-enhanced image. The 3x3 median filter algorithm was used to remove salt and pepper noise and improve image clarity. The Canny edge detection algorithm was applied, and the low threshold was set to 50 and the high threshold was set to 150 to highlight the surface texture features. Morphological operations were performed on the edge-enhanced texture image, and 3x3 structural elements were used for dilation and corrosion to identify and segment possible defect areas. The region growing algorithm was used, and the growth threshold was set to 10 to further refine the boundaries of the defective area. The features of the defective area after segmentation are extracted, including area, perimeter, roundness and grayscale mean. The extracted features are input into the support vector machine classifier, and the radial basis function kernel is used. The penalty parameter C is set to 1.0 to determine whether the defective area is caused by material defects or interference caused by clamping instability. According to the classification results of the support vector machine, the Otsu algorithm is used to automatically calculate the optimal threshold to screen out texture images that are greatly affected by interference factors. The signal-to-noise ratio of the screened texture image is calculated, and the threshold is set to 20dB. The clarity index is calculated using the Laplace operator, and the threshold is set to 0.5 to ensure the representativeness and reliability of the remaining image data. The texture image data that has passed the quality assessment is used as the data basis for the subsequent analysis of the relationship between cutting parameters and machined surface quality.

[0022] S102, extracting key characteristic parameters that can reflect the degree of tool wear from the preprocessed texture image data, including texture orientation angle, texture period and surface roughness parameters, and constructing a multi-dimensional feature vector of the product surface micro-texture.

[0023] The preprocessed texture image data is subjected to a two-dimensional fast Fourier transform to obtain a two-dimensional power spectrum; the main frequency component is extracted from the two-dimensional power spectrum, and the main direction of the power spectrum is calculated according to the main frequency component to determine the texture orientation angle; the texture period is obtained by calculating the main frequency of the power spectrum; the statistical characteristics of the texture image are calculated using the gray level co-occurrence matrix method, wherein a distance parameter and four directions are set, the contrast, entropy, energy and correlation of the four directions are calculated, and the average value is taken as the surface roughness statistical characteristic parameter; the texture image is subjected to a three-layer decomposition to extract the horizontal, vertical and diagonal detail coefficients at different scales; the energy of the detail coefficient at each scale is calculated according to the detail coefficient to obtain the wavelet energy feature; the texture orientation angle, the texture period, the surface roughness statistical characteristic parameter and the wavelet energy feature are combined to construct a feature vector; each feature in the feature vector is normalized to obtain the normalized feature; the normalized features are spliced ​​to form a multi-dimensional feature vector of the micro texture of the product surface.

[0024] Exemplarily, a two-dimensional fast Fourier transform is performed on the preprocessed texture image data, a two-dimensional power spectrum is calculated, and the main frequency component is extracted from the power spectrum. By calculating the main direction of the power spectrum, the texture orientation angle is determined, and the angle reflects the change in cutting direction caused by cutting tool wear. The main frequency of the power spectrum is calculated to obtain the texture period, and the periodic change reflects the periodic change of the machined surface caused by tool wear. The gray level co-occurrence matrix method is used to calculate the statistical characteristics of the texture image, the distance parameter is set to 1 pixel, the direction is 0°, 45°, 90°, and 135°, the contrast, entropy, energy, and correlation in four directions are calculated, and the average value is taken as the statistical characteristic parameter of the surface roughness. These parameters change with the increase in tool wear, reflecting the decrease in surface quality. The texture image is decomposed into three layers using Daubechies wavelet to extract horizontal, vertical, and diagonal detail coefficients at different scales. The energy of the detail coefficient at each scale is calculated to obtain 9 wavelet energy features. The wavelet energy feature is sensitive to the multi-scale changes in surface texture caused by tool wear and can reflect different degrees of tool wear. The extracted texture orientation angle, texture period, four surface roughness statistical features and nine wavelet energy features are combined to construct a 15-dimensional feature vector. Each feature is normalized to the minimum and maximum values ​​so that all eigenvalues ​​are scaled to the range of 0-1. The normalized features are spliced ​​in sequence to form a multi-dimensional feature vector of the product surface microtexture for subsequent tool wear analysis.

[0025] The preprocessed 512x512 pixel texture image data was subjected to a two-dimensional fast Fourier transform to obtain a frequency domain representation. The two-dimensional power spectrum was calculated, and the main frequency component was determined by finding the position of the maximum point in the power spectrum. Assuming that the coordinates of the maximum point are (100, 200), the texture orientation angle θ = arctan (200 / 100) ≈ 63.4°, and the texture period T = 512 / sqrt (100^2 + 200^2) ≈ 2.3 pixels. The increase in θ indicates that the cutting direction deviates, and the decrease in T reflects the intensification of the periodic change of the surface, both of which indicate that the tool wear is aggravated. Then, the gray level co-occurrence matrix method was used to calculate the texture statistical features. The distance parameter was set to 1 pixel, and the directions were 0°, 45°, 90°, and 135°. The contrast, entropy, energy, and correlation of the four directions were calculated. For example, the contrast in the 0° direction is 1.5, the entropy is 4.2, the energy is 0.08, and the correlation is 0.85. The average values ​​of the four directions are taken as the surface roughness statistical features. Subsequently, the Daubechies4 wavelet is used to perform three-layer decomposition of the image, and a total of 10 sub-bands are obtained, namely LL3, LH3, HL3, HH3, LH2, HL2, HH2, LH1, HL1, and HH1. The energy of the nine detail sub-bands, LH3, HL3, HH3, LH2, HL2, HH2, LH1, HL1, and HH1, is calculated, such as the energy of LH3 is 0.05, the energy of HL3 is 0.06, and the energy of HH3 is 0.03, etc., to form a 9-dimensional wavelet energy feature vector. Finally, 1 texture orientation angle, 1 texture period, 4 surface roughness statistical features, and 9 wavelet energy features are combined to construct a 15-dimensional feature vector. Each feature is normalized to the minimum and maximum values, so that all eigenvalues ​​are scaled to the range of 0-1. For example, the normalized texture orientation angle is 0.7, the texture period is 0.4, the contrast is 0.6, the entropy is 0.8, the energy is 0.3, the correlation is 0.9, and the wavelet energy feature is [0.2, 0.3, 0.1, 0.4, 0.5, 0.2, 0.6, 0.7, 0.3]. These normalized features are concatenated in sequence to form the final 15-dimensional feature vector for subsequent tool wear analysis.

[0026] S103. Based on the historical usage data of the tool, including the number of processed products, the cumulative cutting time and the hardness of the processed material, the tool processing surface texture features at different wear stages are collected and analyzed to establish a corresponding relationship between the tool wear degree and the surface texture features.

[0027] According to the historical usage data of the tool, the number of processed products, the cumulative cutting time and the hardness of the processed material are obtained, and the quantitative index of the degree of tool wear is calculated by weighted summation. The surface of the product processed by the tool is microscopically imaged to obtain a surface texture image. The texture orientation angle, texture period and surface roughness characteristic parameters are extracted from the surface texture image, and the trend of the characteristic parameters changing with the degree of wear is recorded. The mapping relationship between the tool wear index and the surface texture characteristic parameters is established, and the radial basis function is selected as the kernel function for model training. For the surface of the newly processed product, the texture characteristic parameters are extracted and substituted into the model to calculate the wear index of the current tool, and the degree of wear of the tool is judged according to the preset wear degree classification standard.

[0028] Exemplarily, a tool wear index is constructed based on the tool history usage data, including the number of processed products, the cumulative cutting time and the hardness of the processed material. The weight of the number of processed products is set to 0.3, the weight of the cumulative cutting time is set to 0.4, and the weight of the hardness of the processed material is set to 0.3. The quantitative index of the degree of tool wear is calculated by weighted summation. The surface of the product processed by tools with different degrees of wear is microscopically imaged using a scanning electron microscope to obtain surface texture images. The scanning electron microscope acceleration voltage is set to 15kV, the working distance is set to 10mm, and the magnification is set to 2000 times. 50 surface texture images of each of the mild, moderate and severe wear stages are collected. Feature parameters such as texture orientation angle, texture period and surface roughness are extracted from the collected surface texture images, and the trend of these parameters changing with the degree of wear is recorded. The support vector regression algorithm is used to establish the mapping relationship between the tool wear index and the surface texture feature parameters. The radial basis function is selected as the kernel function, the penalty factor C is set to 10, and the epsilon parameter is set to 0.1 for model training. The model performance is evaluated using the cross-validation method, and the model with the smallest mean square error is selected as the final model. For the surface of the newly processed product, the texture feature parameters are extracted and substituted into the trained support vector regression model to calculate the wear index of the current tool. The wear degree classification standard is set: the wear index is less than 0.3 for mild wear, 0.3 to 0.7 for moderate wear, and greater than 0.7 for severe wear, and the wear degree of the tool is judged accordingly. Taking a high-speed steel tool as an example, its historical usage data is collected, including the number of processed products is 5000 pieces, the cumulative cutting time is 100 hours, and the average hardness of the processed material is HRC45. After setting the weight coefficient, the tool wear index is calculated to be 0.62. The surface of the product processed by the tool is imaged using a scanning electron microscope, with the acceleration voltage set to 15kV, the working distance to 10mm, and the magnification to 2000 times. 50 surface texture images are collected, and the feature parameters extracted by the image processing algorithm include the average texture orientation angle of 35°, the average texture period of 2.5μm, and the average surface roughness Ra of 1.2μm. These feature parameters and the tool wear index of 0.62 are used as a set of training data. Repeat this process to collect tool data with different degrees of wear, and obtain a total of 150 sets of training data. Use the support vector regression algorithm to establish the mapping relationship, select the radial basis function as the kernel function, set the penalty factor C to 10, and epsilon to 0.1. The model was trained using the 5-fold cross-validation method to obtain the optimal model with a mean square error of 0.025. The surface of the newly processed product was imaged and feature extracted, and the texture orientation angle was 40°, the texture period was 2.8μm, and the surface roughness Ra was 1.5μm. These parameters were input into the trained model, and the wear index of the current tool was calculated to be 0.75. According to the preset grading standard, the tool is judged to be in a state of severe wear, and it is recommended to replace the tool.

[0029] S104, obtaining an estimated value of the current tool wear degree based on the multi-dimensional feature vector of the product surface micro-texture and the corresponding relationship between the tool wear degree and the surface texture characteristics, and comparing it with the wear warning threshold.

[0030] The micro texture image of the product surface is obtained, and the texture orientation angle, texture period, and surface roughness characteristic parameters are extracted from the micro texture image, and the characteristic parameters are standardized to obtain standardized characteristic parameters. A multi-dimensional feature vector is constructed according to the standardized characteristic parameters in a predetermined order. The K nearest neighbor algorithm is used to calculate the Euclidean distance between the multi-dimensional feature vector and all records in the pre-established tool wear degree and surface texture feature correspondence database, and the K records with the smallest distance are selected as the most similar samples. The selected K most similar samples are weighted according to their similarity with the multi-dimensional feature vector, and the similarity weight is calculated using a Gaussian kernel function. The weighted average method is used to calculate the weighted average of the tool wear degree corresponding to the K samples to obtain the wear degree estimation value of the current tool. The tool wear degree estimation value is compared with the preset wear warning threshold value. If the tool wear degree estimation value exceeds the warning threshold value, it is judged that the current tool has reached the warning state.

[0031] Exemplarily, according to the micro-texture image of the product surface, characteristic parameters such as texture orientation angle, texture period, and surface roughness are extracted. The extracted characteristic parameters are standardized and each parameter value is converted to a range of 0-1. The standardized characteristic parameters are combined in a predetermined order to construct a multi-dimensional feature vector. The tool wear degree and surface texture feature correspondence database established in advance are used for matching. The database contains historical processing data, and each record includes a feature vector and a corresponding tool wear degree. The database is updated by regularly collecting new processing data to maintain the timeliness of the data. The K nearest neighbor algorithm is used to calculate the Euclidean distance between the current feature vector and all records in the database, and the K records with the smallest distance are selected as the most similar samples. The K value is set to 5 to balance the calculation efficiency and the accuracy of the result. The selected K most similar samples are weighted according to their similarity with the current feature vector. The similarity weight is calculated using the Gaussian kernel function, and the closer the distance, the greater the weight. The weighted average method is used to calculate the weighted average of the tool wear degree corresponding to the K samples to obtain the wear degree estimate of the current tool. The calculated tool wear degree estimate is compared with the preset wear warning threshold. The warning threshold is determined by analyzing the statistical distribution of historical data and is set to the 75% quantile of the wear degree. If the estimated value exceeds the warning threshold, the current tool is judged to have reached the warning state. According to the degree of exceeding the threshold, the warning is divided into three levels: mild, moderate and severe, corresponding to the situations of exceeding the threshold by 0-10%, 10%-20% and more than 20%, respectively. Taking a high-speed steel milling cutter as an example, the feature parameters extracted from the micro-texture image of the workpiece surface include a texture orientation angle of 35°, a texture period of 2.5μm, and a surface roughness Ra of 1.2μm. These parameters are standardized, and the texture orientation angle is divided by 360° to obtain 0.0972, the texture period is divided by 10μm to obtain 0.25, and the surface roughness is divided by 5μm to obtain 0.24, forming a feature vector [0.0972, 0.25, 0.24]. In a database containing 10,000 historical records, the K nearest neighbor algorithm is used, with K=5 to calculate the Euclidean distance and find the 5 most similar samples. Assume that the feature vectors and corresponding tool wear degrees of these five samples are [0.1, 0.26, 0.23] corresponding to 0.65, [0.09, 0.24, 0.25] corresponding to 0.68,

[0032] [0.11, 0.25, 0.24] corresponds to 0.70, [0.095, 0.27, 0.22] corresponds to 0.66,

[0033] [0.10, 0.23, 0.26] corresponds to 0.69. The Gaussian kernel function is used to calculate the similarity weights, and the weights are 0.25, 0.30, 0.15, 0.20, and 0.10 respectively. The estimated value of the wear degree of the current tool is 0.6725 through weighted average calculation. Historical data analysis shows that the 75% quantile of the wear degree is 0.65, which is set as the warning threshold. The current estimated value of 0.6725 exceeds the threshold of 0.65 by about 9%, which is judged as a mild warning state, prompting close attention to the wear of the tool.

[0034] S105. If the current tool wear level exceeds the preset threshold, a tool replacement warning is triggered, and the batch of products is marked as suspected defective products. The wear critical point is determined by the vibration signal and the cutting force change trend to determine the best time to replace the tool.

[0035] The real-time monitored tool wear degree data is obtained and compared with a preset wear threshold, which is determined by analyzing the statistical distribution of historical processing data; if the tool wear degree exceeds the wear threshold, a tool replacement warning signal is triggered; a tool replacement request is generated based on the tool replacement warning signal; the inventory is automatically checked based on the tool replacement request, and if the inventory is sufficient, the tool preparation process is started; vibration signals and cutting force data are collected and pre-processed in real time; a tool status score is calculated based on the tool wear degree, the vibration signal and the cutting force data, and when the tool status score is lower than the preset threshold, it is determined to be the best time to replace the tool, and the tool replacement process is triggered.

[0036] Exemplarily, the tool wear degree data monitored in real time is compared with the preset wear threshold. The wear threshold is determined by analyzing the statistical distribution of historical processing data and is set to the 75% quantile of the wear degree. If the wear degree exceeds the threshold, a tool replacement warning signal is triggered and the signal is transmitted to the production management system. The production management system automatically generates a tool replacement request and automatically checks the inventory. If the inventory is sufficient, the tool preparation process is started, including tool removal, preset and transportation to the processing area. Product batch tracking is performed, and the current processed product batch is marked as a potential quality defective product, and relevant information is recorded. The quality inspection frequency of the marked batch is increased from the original 1 piece per 100 pieces to 1 piece per 50 pieces, and a special mark is added to the product for subsequent tracking. The vibration sensor and dynamometer installed on the processing equipment are used to collect vibration signals and cutting force data in real time. Fast Fourier transform is performed on the vibration signal to extract the main frequency component and amplitude; wavelet analysis is applied to the cutting force data to extract energy characteristics of different scales. The extracted characteristic parameters are compared with the pre-established wear critical point database to determine whether the tool has reached the wear critical point. The tool status score is calculated using a weighted scoring method by comprehensively considering the degree of wear, vibration signal and cutting force data. When the score is lower than the preset threshold, it is determined to be the best time to replace the tool and the automatic replacement process is triggered. Taking a high-speed steel milling cutter as an example, real-time monitoring shows that its wear degree reaches 0.76. By analyzing historical data, the 75% quantile is determined to be 0.70, which is set as the wear threshold. Since 0.76 exceeds 0.70, the system triggers a tool replacement warning signal and sends it to the production management system. The production management system generates a tool replacement request and sends it to the tool library management module. The tool library automatically checks and finds that the inventory of high-speed steel milling cutters is 5 pieces, which meets the replacement requirements and starts the preparation process. At the same time, the product batch tracking system marks the 100 products currently being processed as potential quality defects, records the information in the quality control database, and increases the inspection frequency to 1 piece for every 50 pieces. After the signal collected by the vibration sensor is fast Fourier transformed, the main frequency is 1200Hz and the amplitude is 0.5g. After wavelet analysis, the cutting force data collected by the dynamometer has an energy eigenvalue of 0.8 in the third layer of wavelet decomposition. Comparing these features with the wear critical point database, it is found that the vibration amplitude exceeds the critical value of 0.4g and the cutting force energy feature is close to the critical value of 0.85. Using the weighted scoring method, the wear degree weight is set to 0.4, the vibration signal weight is set to 0.3, and the cutting force data weight is set to 0.3. The tool status score is calculated to be 68 points, with a full score of 100 points. This score is lower than the preset threshold of 70 points, and it is determined that the best replacement time has been reached, and the tool replacement process is automatically triggered.

[0037] Vibration sensors and cutting force sensors are used to monitor the machining process in real time, extract the characteristic frequency and amplitude of vibration and cutting force, and establish the corresponding relationship between the vibration amplitude, spindle current, cutting force and the degree of tool wear. When the vibration amplitude suddenly increases and the cutting force continues to rise and exceeds the preset threshold, it is determined that the tool has reached the critical point of wear.

[0038] The vibration signal and cutting force signal collected in real time during the machining process are de-noised to obtain the denoised vibration signal and cutting force signal. According to the denoised vibration signal and cutting force signal, the characteristic frequency and amplitude of the vibration and cutting force are extracted by fast Fourier transform. The spindle current value is collected by using a Hall current sensor, and the stable current data is obtained by low-pass filtering. The corresponding relationship between the tool wear degree and the vibration amplitude, spindle current and cutting force is established, and the model is trained by using the support vector regression algorithm according to the corresponding relationship data. The vibration amplitude, spindle current and cutting force data are input into the model to determine the corresponding tool wear degree estimation value. The warning threshold of the vibration amplitude and cutting force is set, and the warning threshold is determined by using a multivariate statistical analysis method according to the tool type and the processing material. The current vibration amplitude and cutting force data are compared with the preset threshold in real time, and the current tool wear degree is estimated by using the trained support vector regression model. The weighted scoring method is used to set the vibration amplitude weight, cutting force weight and estimated wear degree weight, and calculate the comprehensive score. If the comprehensive score exceeds a preset critical value, or the vibration amplitude exceeds a sudden increase threshold, or the continuous measurement value of the cutting force exceeds a continuous increase threshold, it is determined that the tool has reached a critical point of wear.

[0039] Exemplarily, vibration sensors and cutting force sensors are used to collect vibration signals and cutting force signals in the processing process in real time, bandpass filters are used to remove noise, and fast Fourier transform is used to extract the characteristic frequency and amplitude of vibration and cutting force. The spindle current value is collected by Hall current sensor, and stable current data is obtained after low-pass filtering. A database of the corresponding relationship between tool wear degree and vibration amplitude, spindle current and cutting force is established, and the model is trained using support vector regression algorithm. The radial basis function is selected as the kernel function, the regularization parameter C is set to 1.0, the epsilon parameter is set to 0.1, the vibration amplitude, spindle current and cutting force data are input, and the corresponding tool wear degree estimation value is output. The warning threshold of vibration amplitude and cutting force is set, and the threshold is determined by multivariate statistical analysis method in combination with tool type and processing material. For high-speed steel tool processing 45 steel, the 90% quantile of vibration amplitude is used as the sudden increase threshold, and the continuous increase is defined when the cutting force measurement values ​​exceed 85% quantile for 5 consecutive times. The current vibration amplitude and cutting force data are compared with the preset threshold in real time, and the current tool wear degree is estimated by the trained support vector regression model. The weighted scoring method is used to set the vibration amplitude weight to 0.3, the cutting force weight to 0.3, and the estimated wear degree weight to 0.4 to calculate the comprehensive score. If the comprehensive score exceeds the preset critical value, or the vibration amplitude exceeds the sudden increase threshold, or the cutting force measurement value exceeds the continuous increase threshold for 5 consecutive times, it is determined that the tool has reached the critical point of wear and the tool replacement process is triggered. Taking the high-speed steel milling cutter for processing 45 steel as an example, the original signal collected by the vibration sensor is filtered by 50-5000Hz bandpass, and the main frequency is 1200Hz and the amplitude is 0.5g through fast Fourier transform. The signal collected by the cutting force sensor is filtered by 100-2000Hz bandpass, with a main frequency of 800Hz and an amplitude of 500N. The spindle current collected by the Hall current sensor is stabilized at 15A after 20Hz low-pass filtering. These data are input into the trained support vector regression model, whose kernel function is RBF, C=1.0, epsilon=0.1, and the current tool wear degree is estimated to be 0.68. According to historical data statistics, the threshold for sudden increase in vibration amplitude is set to 0.8g (90% quantile), and the threshold for continuous increase in cutting force is set to 550N (85% quantile). Real-time data shows that the vibration amplitude of 0.5g does not exceed the threshold, but the cutting force is measured for 5 consecutive times, namely 520N, 530N, 535N, 542N, and 551N, and the last time exceeds the threshold. Using the weighted scoring method, the vibration amplitude score is (0.5 / 0.8)*30=18.75, the cutting force score is (551 / 550)*30=30.05, the estimated wear degree score is (0.68 / 0.75)*40=36.27, and the comprehensive score is 85.07. The critical value of the comprehensive score is set to 80, and the current score exceeds the critical value. It is comprehensively judged that the tool has reached the critical point of wear, and the tool replacement process should be automatically triggered.

[0040] S106. For products suspected of being defective, the defect type is identified by analyzing the surface texture, and the material properties and design dimensional tolerance factors are comprehensively analyzed to determine whether the abnormal surface texture is caused by tool wear or a defect in the product itself.

[0041] Perform a surface scan on the identified suspected defective products, and use a grayscale co-occurrence matrix to extract texture features, the texture features including contrast, entropy, energy and correlation, and extract local texture information through local binary patterns; construct a surface texture feature vector based on the texture features, classify the surface texture feature vectors, and obtain preliminary defect type recognition results and their probability distribution; obtain the material property data of the product, and at the same time obtain the dimensional tolerance requirements in the design drawings; use the material property data, the dimensional tolerance requirements and the current processing parameters to establish a material, tolerance and process feature vector; if the material, tolerance and process feature vector is obtained, perform a comprehensive analysis based on the defect type recognition results and their probability distribution; determine whether the surface texture anomaly is caused by tool wear or a defect in the product itself through the comprehensive analysis, and output the confidence level of the judgment result.

[0042] Exemplarily, a high-resolution surface scan is performed on the identified suspected defective products, and the gray-level co-occurrence matrix is ​​used to extract texture features, including contrast, entropy, energy, and correlation. Local binary patterns are used to extract local texture information, and multi-scale texture features are obtained through wavelet transform. All extracted features are normalized to construct a comprehensive surface texture feature vector. The surface texture feature vector is classified using the support vector machine algorithm, the radial basis function is selected as the kernel function, and the penalty factor C is set to 1.0. The texture anomalies are divided into two categories: wear and defect, and the preliminary defect type identification results and their probability distribution are obtained. The material property data of the product, including hardness, strength, and thermal conductivity, are extracted from the product database. At the same time, the dimensional tolerance requirements in the design drawings are obtained, and the current processing parameters, such as cutting speed, feed rate, and cutting depth, are added to establish the material and tolerance and process feature vectors. Combined with the surface texture classification results and their probability distribution, the material and tolerance and process feature vectors, the random forest algorithm is used for comprehensive analysis. The number of decision trees was set to 100, and the maximum tree depth was set to 10. The majority voting principle was used to determine whether the surface texture abnormality was caused by tool wear or product defects, and the confidence of the judgment result was output. Taking the milling surface of a 45# steel as an example, a three-dimensional surface scanner with a resolution of 0.1μm was used to obtain the surface morphology data. The gray-level co-occurrence matrix was used to calculate the contrast of 1.5, entropy of 4.2, energy of 0.08, and correlation of 0.85. The LBP algorithm was used to extract local texture features and obtain a 256-dimensional feature vector. Multi-scale texture features were obtained through 5-layer wavelet decomposition. All features were normalized to the range of 0-1, and a 768-dimensional comprehensive surface texture feature vector was constructed. The support vector machine with RBF kernel function was used for classification, C=1.0, and the preliminary classification result was wear class with a probability of 0.75. The material properties extracted from the database include hardness HRC45, tensile strength 600MPa, thermal conductivity 50.2W / (m·K); the design tolerance requirement is flatness 0.02mm, and the current processing parameters are cutting speed 100m / min, feed rate 0.1mm / r, and cutting depth 0.5mm. These data are combined into a 15-dimensional material, tolerance, and process feature vector. A random forest algorithm is used with 100 trees and a maximum depth of 10 to comprehensively analyze all features. After majority voting, 75 decision trees are judged to be caused by tool wear, and 25 are judged to be defects of the product itself. The final output result is a surface anomaly caused by tool wear, with a confidence level of 75%.

[0043] Combined with the product CAD model, key dimensions and tolerance requirements are extracted, and a correlation model of material properties, dimensional tolerances and surface texture characteristics is established. If the product surface texture abnormality is consistent with the material properties and tolerance requirements, it is judged to be a defect of the product itself; if the degree of abnormality exceeds the allowable range of the material and tolerance, and is consistent with the tool wear characteristics, it is judged to be caused by tool wear.

[0044] Extract key dimensions and tolerance requirements from the product CAD model to obtain the product's geometric features, dimensional values ​​and tolerance ranges. Use the support vector regression method to establish a correlation model between material properties, dimensional tolerances and surface texture features. According to the correlation model, input the standardized material properties and tolerance data to obtain the expected surface texture parameter range. Compare the surface texture anomaly with the prediction results of the correlation model, and set the judgment threshold to the standard deviation multiple of the prediction range. If the degree of anomaly is within the threshold, it is judged as a defect of the product itself; if it exceeds the threshold, it is matched with the tool wear feature library. Use cosine similarity to calculate the similarity between the anomaly and the features in the library to determine whether the anomaly is caused by tool wear.

[0045] Exemplarily, a material property database is established, and the hardness, toughness and thermal conductivity parameters of different materials are entered. The data are stored in a structured data format, and an index is set to improve query efficiency. Material property data is obtained through standardized experimental tests, and literature data is collected for cross-validation to ensure data accuracy. Key dimensions and tolerance requirements are extracted from the product CAD model, and the main geometric features of the product are identified using a rule-based feature recognition algorithm to obtain its dimensional value and tolerance range. For complex shapes, a convolutional neural network is used for feature extraction and classification. A correlation model of material properties, dimensional tolerances and surface texture features is established using the support vector regression method. The radial basis kernel function is selected, the penalty factor C is set to 1.0, the epsilon parameter is set to 0.1, the standardized material properties and tolerance data are input, and the expected surface texture parameter range is output. The model performance is evaluated using K-fold cross validation, and the model with the smallest mean square error is selected as the final model. The observed surface texture anomaly is compared with the prediction results of the correlation model, and the judgment threshold is set to 1.5 times the standard deviation of the prediction range. If the degree of anomaly is within the threshold, it is determined to be a defect of the product itself; if it exceeds the threshold, it is matched with the tool wear feature library. The tool wear feature library is established through historical data statistics and expert knowledge and is updated regularly. The cosine similarity is used to calculate the similarity between the observed anomaly and the features in the library. If the similarity exceeds 0.8, it is determined to be caused by tool wear. Taking a 45# steel part as an example, its hardness HRC45, toughness 100J / cm2, and thermal conductivity 50.2W / (m·K) are obtained from the material property database. The key dimensions are extracted by the CAD model feature recognition algorithm: diameter 50±0.02mm, length 100±0.05mm. These data are input into the support vector regression model, C=1.0, epsilon=0.1, and the predicted surface roughness Ra range is 1.2~1.8μm. After actual processing, the precision roughness meter is used to measure Ra=2.1μm, which exceeds the predicted range by 1.5 times the standard deviation (0.3μm). It is automatically matched with the tool wear feature library, which contains 100 sets of historical wear data. The maximum cosine similarity between the current anomaly and the features in the library is calculated to be 0.85, which exceeds the threshold of 0.8. The comprehensive judgment result is that the surface anomaly is caused by tool wear, not a defect in the product itself. The tool replacement process is triggered, and the data is added to the wear feature library for subsequent analysis and prediction.

[0046] S107. Based on the judgment results, the product surface texture detection results and the tool wear warning information are correlated and analyzed to build a two-way feedback of product quality prediction and tool life optimization. By continuously collecting production site data, the correspondence between the tool wear degree and the surface texture characteristics is dynamically updated and optimized.

[0047] Obtain product surface texture detection results and tool wear warning information; measure the degree of tool wear to obtain tool wear data. According to the tool wear data and processing parameters, establish a product quality prediction model, and the product quality prediction model selects radial basis function as the kernel function. The output result of the product quality prediction model is compared with the preset quality standard to determine whether the product is qualified. If the product is unqualified, optimize the tool replacement timing based on the genetic algorithm, and the genetic algorithm sets the chromosome encoding to represent different tool replacement strategies. According to the optimization result of the tool replacement strategy, an online learning algorithm is used to realize incremental learning. The online learning algorithm continuously collects production site data and updates the parameters of the product quality prediction model.

[0048] Exemplarily, the association data between the product surface texture detection results and the tool wear warning information is constructed, and the time series data storage structure is used to record the surface texture parameters and the corresponding tool wear degree during each processing. The tool wear degree is measured by combining optical sensors and vibration signal analysis to achieve real-time monitoring. The product quality prediction model is established using the support vector regression algorithm, the radial basis function is selected as the kernel function, the penalty factor C is set to 1.0, the epsilon is set to 0.1, the tool wear degree and processing parameters are input, and the predicted product surface texture parameters are output. The prediction results are compared with the preset quality standards to determine whether the product is qualified, and the results are fed back to the tool replacement strategy optimization module. The tool replacement timing is optimized based on the genetic algorithm, the chromosome encoding is set to represent different tool replacement strategies, and new strategies are generated through crossover and mutation operations. A multi-objective fitness function is constructed, including product qualification rate, tool cost, equipment utilization rate and product delivery time, and the weight coefficients are set to 0.4, 0.3, 0.2 and 0.1 respectively to balance product quality and tool cost. Iterative optimization obtains the best replacement timing, and the results are applied to production scheduling. An online learning algorithm is used to achieve incremental learning. Production site data is continuously collected, and the parameters of the product quality prediction model and tool life optimization algorithm are updated every 100 new samples. The sliding window method is used to evaluate the performance of the updated model. If the performance drops by more than 5%, the model retraining process is triggered. The corresponding relationship between the tool wear degree and the surface texture characteristics is dynamically adjusted to achieve adaptive optimization of the model. Taking a high-speed steel milling cutter processing 45 steel as an example, the associated database records the data of nearly 1000 processing processes. The optical sensor measures the tool wear amount as 0.15mm, and the vibration signal analysis obtains the main frequency of 1200Hz and the amplitude of 0.5g. The comprehensive judgment of the tool wear degree is 0.68. These data are input into the support vector regression model, C=1.0, epsilon=0.1, and the predicted product surface roughness Ra is 1.8μm, which exceeds the preset quality standard of 1.6μm by 12.5%. The genetic algorithm optimizes the tool replacement strategy, the population size is set to 100, and the iteration is 50 generations. The fitness function calculation results are: product qualification rate 95% (weight 0.4), tool cost reduction 8% (weight 0.3), equipment utilization rate increase 5% (weight 0.2), product delivery time reduction 2%

[0049] (weight 0.1), the comprehensive score is 0.856. After optimization, it is recommended to replace the tool when the wear degree reaches 0.75. After collecting 100 new samples, the online learning algorithm updates the model parameters. The sliding window method, where the window size is 200, evaluates the performance of the updated model, and the mean square error is reduced from 0.025 to 0.022, and the performance is improved by 12%. The corresponding relationship between the tool wear degree and the surface texture characteristics is automatically adjusted, and the weight of the wear degree in the roughness Ra prediction model is increased from 0.6 to 0.65, reflecting the enhanced influence of the tool status on the processing quality.

[0050] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the concept of the present application. For example, the above features are replaced with the technical features with similar functions disclosed in the present application (but not limited to) to form a technical solution.

Claims

1. A production line product quality traceability method based on industrial vision, characterized in that: The method comprises: Use imaging equipment to perform microscopic imaging of the product surface, obtain texture image data of the machined surface under different cutting parameters, perform image preprocessing to improve image contrast and clarity, and eliminate interference factors such as material defects, clamping instability, and product defects themselves; From the preprocessed texture image data, key characteristic parameters that can reflect the degree of tool wear are extracted, including texture orientation angle, texture period and surface roughness parameters, and a multi-dimensional feature vector of the product surface micro-texture is constructed; Based on the historical usage data of the tool, including the number of processed products, cumulative cutting time and hardness of the processed material, the surface texture characteristics of the tool processing at different wear stages are collected and analyzed to establish the corresponding relationship between the degree of tool wear and the surface texture characteristics; According to the multi-dimensional feature vector of the micro-texture of the product surface, combined with the correspondence between the tool wear degree and the surface texture characteristics, the estimated value of the current tool wear degree is obtained and compared with the wear warning threshold; If the current tool wear exceeds the preset threshold, a tool replacement warning is triggered, and the batch of products is marked as suspected defective products. The wear critical point is determined by the vibration signal and cutting force change trend to determine the best time to replace the tool. For products suspected of defects that have been identified, the defect type is identified by analyzing the surface texture, and the material properties and design dimensional tolerance factors are comprehensively analyzed to determine whether the abnormal surface texture is caused by tool wear or product defects. Based on the judgment results, the product surface texture detection results and tool wear warning information are correlated and analyzed to build a two-way feedback loop for product quality prediction and tool life optimization. By continuously collecting production site data, the correspondence between tool wear degree and surface texture characteristics is dynamically updated and optimized.

2. The method according to claim 1, wherein: The imaging device is used to perform microscopic imaging of the product surface, obtain texture image data of the machined surface under different cutting parameters, perform image preprocessing to improve image contrast and clarity, and eliminate interference factors such as material defects, clamping instability, and product defects themselves, including: Use imaging equipment to perform microscopic imaging on the product surface to obtain texture image data of the product surface; Preprocessing the texture image data to adjust the pixel grayscale distribution of the texture image data; If the texture image data has noise, removing the noise; Highlighting the surface texture features of the texture image data according to the Canny edge detection algorithm to obtain an edge-enhanced texture image; Performing morphological operations on the edge-enhanced texture image, including dilation and erosion, to identify and segment defective areas in the edge-enhanced texture image; Extract features of defect areas, including area, perimeter, roundness, and grayscale statistics; Input the features of the defect area into the support vector machine classifier to determine whether the defect area is caused by material defects or clamping instability; According to the judgment results of the support vector machine classifier, the texture images that are greatly affected by interference factors are screened out.

3. The method according to claim 1, wherein: The preprocessed texture image data is used to extract key characteristic parameters that can reflect the degree of tool wear, including texture orientation angle, texture period and surface roughness parameters, and to construct a multi-dimensional feature vector of the product surface micro-texture, including: Performing a two-dimensional fast Fourier transform on the preprocessed texture image data to obtain a two-dimensional power spectrum; Extracting a main frequency component from the two-dimensional power spectrum, calculating a main direction of the power spectrum according to the main frequency component, and determining a texture orientation angle; The texture period is obtained by calculating the main frequency of the power spectrum; The gray-level co-occurrence matrix method is used to calculate the statistical characteristics of the texture image, wherein the distance parameter and four directions are set, the contrast, entropy, energy and correlation of the four directions are calculated, and the average value is taken as the statistical characteristic parameter of the surface roughness; The texture image is decomposed into three layers to extract horizontal, vertical and diagonal detail coefficients at different scales; Calculate the energy of detail coefficients at each scale according to the detail coefficients to obtain wavelet energy features; Combining the texture orientation angle, the texture period, the surface roughness statistical characteristic parameter and the wavelet energy feature to construct a feature vector; Normalizing each feature in the feature vector to obtain a normalized feature; The normalized features are concatenated to form a multi-dimensional feature vector of the micro texture of the product surface.

4. The method according to claim 1, wherein: According to the tool's historical usage data, including the number of processed products, cumulative cutting time, and hardness of processed materials, the tool processing surface texture features at different wear stages are collected and analyzed, and the corresponding relationship between the tool wear degree and the surface texture features is established, including: According to the tool's historical usage data, the number of processed products, the cumulative cutting time and the hardness of the processed material are obtained, and the quantitative index of the tool wear degree is obtained through weighted summation calculation; Microscopic imaging of the surface of the product processed by the tool to obtain the surface texture image; Extracting texture orientation angle, texture period and surface roughness characteristic parameters from the surface texture image, and recording the trend of the characteristic parameters changing with the degree of wear; The mapping relationship between tool wear index and surface texture characteristic parameters is established, and radial basis function is selected as the kernel function for model training; For the surface of a newly processed product, texture feature parameters are extracted and substituted into the model to calculate the wear index of the current tool, and the wear degree of the tool is judged according to the preset wear degree classification standard.

5. The method according to claim 1, wherein: The method of obtaining an estimated value of the current tool wear degree based on the multi-dimensional feature vector of the product surface micro-texture and the corresponding relationship between the tool wear degree and the surface texture feature, and comparing it with the wear warning threshold, includes: Acquire a micro texture image of the product surface, extract texture orientation angle, texture period and surface roughness characteristic parameters from the micro texture image, and perform standardization on the characteristic parameters to obtain standardized characteristic parameters; Constructing a multi-dimensional feature vector according to the standardized feature parameters combined in a predetermined order; The K nearest neighbor algorithm is used to calculate the Euclidean distance between the multi-dimensional feature vector and all records in the pre-established database of correspondence between tool wear degree and surface texture features, and the K records with the smallest distance are selected as the most similar samples; The selected K most similar samples are weighted according to their similarity with the multi-dimensional feature vector, and the similarity weight is calculated using a Gaussian kernel function; The weighted average method is used to calculate the weighted average of the tool wear degrees corresponding to the K samples to obtain the estimated value of the current tool wear degree; The tool wear degree estimation value is compared with a preset wear warning threshold value. If the tool wear degree estimation value exceeds the warning threshold value, it is determined that the current tool reaches a warning state.

6. The method according to claim 1, wherein: If the current tool wear exceeds the preset threshold, a tool replacement warning is triggered, and the batch of products is marked as suspected defective products. The wear critical point is determined by the vibration signal and the cutting force change trend, and the best time to replace the tool is determined, including: Acquire real-time monitored tool wear degree data and compare it with a preset wear threshold value, wherein the wear threshold value is determined by analyzing the statistical distribution of historical processing data; If the wear degree of the tool exceeds the wear threshold, a tool replacement warning signal is triggered; generating a tool change request according to the tool change warning signal; Automatically check inventory based on the tool change request, and if the inventory is sufficient, start the tool preparation process; Collect and pre-process vibration signals and cutting force data in real time; Calculating a tool status score according to the tool wear degree, the vibration signal and the cutting force data, and determining that it is the best time to replace the tool when the tool status score is lower than a preset threshold, and triggering a tool replacement process; It also includes: using vibration sensors and cutting force sensors to monitor the machining process in real time, extracting the characteristic frequency and amplitude of vibration and cutting force, establishing the corresponding relationship between the vibration amplitude, spindle current and cutting force and the degree of tool wear, and when the vibration amplitude suddenly increases and the cutting force continues to rise and exceeds the preset threshold, it is determined that the tool has reached the critical point of wear.

7. The method according to claim 6, wherein: The method uses a vibration sensor and a cutting force sensor to monitor the machining process in real time, extracts the characteristic frequency and amplitude of vibration and cutting force, establishes a corresponding relationship between the vibration amplitude, spindle current and cutting force and the degree of tool wear, and determines that the tool has reached a critical point of wear when the vibration amplitude suddenly increases and the cutting force continues to rise and exceeds a preset threshold, including: De-noising the vibration signal and cutting force signal collected in real time during the machining process to obtain the de-noised vibration signal and cutting force signal; Extracting the characteristic frequency and amplitude of the vibration and cutting force by fast Fourier transform according to the denoised vibration signal and cutting force signal; The spindle current value is collected by using the Hall current sensor, and stable current data is obtained through low-pass filtering. Establishing the corresponding relationship between tool wear degree and vibration amplitude, spindle current and cutting force, and training the model using support vector regression algorithm based on the corresponding relationship data; Input vibration amplitude, spindle current and cutting force data into the model to determine the corresponding tool wear degree estimate; Setting warning thresholds for vibration amplitude and cutting force, and determining the warning thresholds using a multivariate statistical analysis method based on tool type and processing material; Compare the current vibration amplitude, cutting force data and the preset threshold in real time, and estimate the current tool wear degree using the trained support vector regression model; The weighted scoring method is used to set the vibration amplitude weight, cutting force weight and estimated wear degree weight to calculate the comprehensive score; If the comprehensive score exceeds a preset critical value, or the vibration amplitude exceeds a sudden increase threshold, or the continuous measurement value of the cutting force exceeds a continuous increase threshold, it is determined that the tool has reached a critical point of wear.

8. The method according to claim 1, wherein: For the suspected defective products that have been marked, the defect type is identified by analyzing the surface texture, and the material properties and design dimensional tolerance factors are comprehensively analyzed to determine whether the abnormal surface texture is caused by tool wear or product defects, including: Scan the surface of the identified suspected defective products and extract texture features using a gray-level co-occurrence matrix, wherein the texture features include contrast, entropy, energy, and correlation, and extract local texture information through a local binary pattern; Constructing a surface texture feature vector according to the texture feature, classifying the surface texture feature vector, and obtaining a preliminary defect type recognition result and its probability distribution; Obtain the material property data of the product and the dimensional tolerance requirements in the design drawings; Establishing a material and tolerance and process feature vector using the material property data, the dimensional tolerance requirements and current processing parameters; If the material, tolerance and process feature vectors are obtained, a comprehensive analysis is performed in combination with the defect type identification results and their probability distribution; Through the comprehensive analysis, it is determined whether the abnormal surface texture is caused by tool wear or a defect in the product itself, and the confidence level of the determination result is output; It also includes: combining the product CAD model, extracting key dimensions and tolerance requirements, and establishing a correlation model between material properties, dimensional tolerances and surface texture characteristics. If the product surface texture abnormality is consistent with the material properties and tolerance requirements, it is judged to be a defect of the product itself; if the degree of abnormality exceeds the allowable range of the material and tolerance, and is consistent with the tool wear characteristics, it is judged to be caused by tool wear.

9. The method according to claim 8, wherein: The method combines the product CAD model to extract key dimensions and tolerance requirements, establishes a correlation model between material properties, dimensional tolerances and surface texture features, and determines that the product itself is defective if the surface texture abnormality of the product is consistent with the material properties and tolerance requirements; If the degree of abnormality exceeds the allowable range of materials and tolerances and is consistent with the characteristics of tool wear, it is determined to be caused by tool wear, including: Extract key dimensions and tolerance requirements from product CAD models to obtain product geometric features, dimension values ​​and tolerance ranges; Using support vector regression method, a correlation model between material properties, dimensional tolerance and surface texture characteristics is established; According to the correlation model, inputting the standardized material properties and tolerance data, and obtaining the expected surface texture parameter range; Compare the surface texture anomaly with the prediction result of the correlation model, and set the judgment threshold to a multiple of the standard deviation of the prediction range; If the degree of abnormality is within the threshold, it is determined to be a defect of the product itself; If it exceeds the threshold, matching is performed with the tool wear feature library; The cosine similarity is used to calculate the similarity between the anomaly and the features in the library to determine whether the anomaly is caused by tool wear.

10. The method according to claim 1, wherein: According to the judgment result, the product surface texture detection result and the tool wear warning information are correlated and analyzed to build a two-way feedback of product quality prediction and tool life optimization. By continuously collecting production site data, the corresponding relationship between tool wear degree and surface texture characteristics is dynamically updated and optimized, including: Obtain product surface texture detection results and tool wear warning information; Measure the tool wear degree and obtain tool wear data; According to the tool wear data and processing parameters, a product quality prediction model is established, wherein the product quality prediction model selects a radial basis function as a kernel function; Compare the output result of the product quality prediction model with the preset quality standard to determine whether the product is qualified; If the product is unqualified, the tool replacement timing is optimized based on a genetic algorithm, where the genetic algorithm sets chromosome codes to represent different tool replacement strategies; According to the optimization result of the tool replacement strategy, an online learning algorithm is used to implement incremental learning. The online learning algorithm continuously collects production site data and updates the parameters of the product quality prediction model.

Images