Intelligent detection method and system for workpiece machining quality of numerical control machine tool

Abstract

The invention relates to the technical field of image data processing, in particular to a numerical control machine tool workpiece machining quality intelligent detection method and system, and the method comprises the steps: collecting the point cloud data of a to-be-detected numerical control machine tool workpiece and a standard numerical control machine tool workpiece, and carrying out the clustering of the point cloud data of the to-be-detected numerical control machine tool workpiece through an improved mean shift algorithm, obtaining a plurality of clusters; the DB index of each cluster is obtained, the clusters with the DB indexes smaller than a set threshold value are removed, and final point cloud data are obtained; and splicing the final point cloud data with the point cloud data of the standard numerical control machine tool workpiece to generate a point cloud data model. Boolean operation is carried out on the point cloud data models of the to-be-detected numerical control machine tool workpiece and the standard numerical control machine tool workpiece, and an abnormal area is identified; and if the area of the abnormal region is greater than a set threshold value, determining that the to-be-detected workpiece is abnormal. The problem that the workpiece machining quality detection accuracy is not high is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing. More specifically, the present invention relates to an intelligent detection method and system for the machining quality of workpieces in a numerically controlled machine tool. Background Art

[0002] With the continuous improvement of the manufacturing industry's requirements for product quality, numerically controlled machine tools are widely used in modern industrial production. Numerically controlled machine tools can achieve high-precision and high-efficiency machining by computer numerical control of the machining process. However, with the increasing complexity of machining processes and the diversification of workpiece shapes, traditional quality inspection methods face many challenges. Traditional methods for inspecting the machining quality of workpieces mainly rely on manual visual inspection, measuring tools, etc. This not only has low efficiency but is also easily affected by human factors, making it difficult to ensure the accuracy and consistency of inspection results.

[0003] In order to improve the accuracy and efficiency of machining quality inspection, in recent years, intelligent detection technologies have gradually been applied to the workpiece quality inspection of numerically controlled machine tools. Traditional intelligent detection methods mostly rely on traditional machine learning models based on features, such as support vector machines, decision trees, etc., for defect identification. However, these methods usually require a large amount of labeled data for training, and are prone to problems such as insufficient detection accuracy and poor adaptability when facing complex workpieces, different machining environments, and dynamic changes. Most of the existing technologies cannot effectively overcome the limitations of traditional models, especially in the identification of complex workpiece shapes, minor changes during the machining process, and new types of defects, and there is still much room for improvement.

[0004] The patent application document with the application publication number CN113866165A discloses a workpiece inspection and defect detection system including monitoring workpiece images. This patent application document uses the workpiece images obtained by a camera as training data to train a defect detection model to identify workpiece images containing defects. Based on the features of these training images, the classification characteristics of an anomaly detector are further determined, and the determined classification characteristics of the anomaly detector are used to determine whether the workpiece image belongs to an abnormal category.

[0005] However, the above technical solutions mainly rely on traditional models to achieve intelligent detection of workpieces. These models are usually trained based on existing standard features and algorithms, without fully considering the limitations of existing models, resulting in low accuracy in detecting the machining quality of workpieces when facing complex workpieces or changes in the machining environment. Summary of the Invention

[0006] To solve the problem of low accuracy in detecting the machining quality of workpieces proposed in the above background art, the present invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides an intelligent detection method for the machining quality of a numerically controlled machine tool workpiece, including: collecting point cloud data of each position of the workpiece to be detected on the numerically controlled machine tool and the standard numerically controlled machine tool workpiece; using an improved mean shift algorithm to cluster the point cloud data of each position of the workpiece to be detected on the numerically controlled machine tool to obtain a plurality of clustering clusters; obtaining the DB index of each clustering cluster, and if the DB index is less than a first set threshold, removing the corresponding cluster to obtain the final point cloud data; splicing the final point cloud data and the point cloud data of each position of the standard numerically controlled machine tool workpiece to obtain a point cloud data model of the workpiece to be detected on the numerically controlled machine tool and the standard numerically controlled machine tool workpiece; performing a Boolean operation on the point cloud data models of the workpiece to be detected on the numerically controlled machine tool and the standard numerically controlled machine tool workpiece to obtain the area of the abnormal region, and if the area of the abnormal region is greater than a set area threshold, determining that the workpiece to be detected on the numerically controlled machine tool is abnormal; wherein, the improved mean shift algorithm includes a drift vector, and calculates the drift vector of the th iteration of the th data point belonging to the th clustering cluster , , where is the initial drift vector of the th iteration of the th data point belonging to the th clustering cluster, is the DB index of the th iteration of the th data point belonging to the th clustering cluster, is the DB index of the th iteration of the th data point belonging to the th clustering cluster, is an empirical constant.

[0008] The above technical solution clusters the point cloud data through an improved mean shift algorithm, removes noise and screens out potential abnormal clusters. The ICP algorithm is used to splice the point cloud data of the workpiece to be detected and the standard workpiece to generate an accurate three-dimensional model. The area of the abnormal region is calculated through a Boolean operation, and combined with the area threshold to determine whether it is an abnormal workpiece, thus realizing efficient and accurate machining quality detection and improving the accuracy and automation level of detection.

[0009] Further, the obtaining of the DB index of each clustering cluster is specifically: Using the mean method to obtain the cluster center of each clustering cluster, calculating the inter-cluster distance and intra-cluster distance of each clustering cluster; taking the mean of the inter-cluster distance and intra-cluster distance of each clustering cluster as the DB index of the clustering cluster.

[0010] The above technical solution obtains the cluster centers of the clustering clusters through the mean method, and calculates the mean of the inter-cluster distance and the intra-cluster distance as the DB index, effectively improving the accuracy of the clustering result. This method can accurately reflect the density of each clustering cluster and the degree of separation from other clusters, making the identification of abnormal clusters more reliable. Through the comprehensive evaluation of the DB index, data clusters with abnormal characteristics can be accurately screened out, further improving the effect of clustering analysis and providing more stable and accurate data support for subsequent defect detection.

[0011] Furthermore, a binocular camera or a structured light camera is used to collect the point cloud data of each position of the workpiece of the numerically controlled machine tool to be detected and the standard numerically controlled machine tool workpiece.

[0012] The above technical solution can achieve high-precision three-dimensional data acquisition by using a binocular camera or a structured light camera to collect the point cloud data of the workpiece of the numerically controlled machine tool to be detected and the standard numerically controlled machine tool workpiece. These cameras can provide detailed depth information and rich geometric features, effectively capturing the details of the workpiece surface, including tiny shape variations or defects. Compared with traditional two-dimensional images, the point cloud data can comprehensively reflect the three-dimensional structure of the workpiece, making subsequent detection, comparison and analysis more accurate, thus significantly improving the accuracy and reliability of the detection.

[0013] Furthermore, it also includes performing mean filtering processing on the point cloud data.

[0014] The above technical solution can effectively remove noise and outliers, smooth the data, and improve the quality and accuracy of the point cloud by performing mean filtering processing on the point cloud data. This preprocessing method reduces the unstable factors caused by measurement errors or external interferences, making the subsequent clustering, registration and defect detection processes more accurate and reliable. The application of mean filtering improves the consistency and usability of the point cloud data, thereby improving the robustness and detection accuracy of the entire detection.

[0015] Furthermore, the ICP algorithm is used to splice the final point cloud data and the point cloud data of each position of the standard numerically controlled machine tool workpiece.

[0016] Furthermore, the set area threshold is 0.5.

[0017] Furthermore, it also includes performing secondary clustering on the point cloud data within the clustering cluster corresponding to the DB index being greater than the second set threshold and less than the first set threshold.

[0018] Through the secondary clustering process for the clustering clusters with the DB index between two set thresholds, the above technical solution can further refine the clustering results and enhance the detection ability of abnormal regions. The secondary clustering can effectively identify the subtle differences in these clustering clusters, avoid missing potential defects, especially when there are complex structures or irregular shapes inside the clustering clusters, and improve the clustering accuracy. This step optimizes the data division, enabling the detection to more accurately capture the minute abnormalities in the workpiece and improving the overall detection sensitivity and reliability.

[0019] In a second aspect, the present invention provides an intelligent detection system for the machining quality of CNC machine tool workpieces, including a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent detection method for the machining quality of CNC machine tool workpieces as described in any one of the above is implemented.

[0020] The beneficial effects of the present invention are as follows: By combining point cloud data processing and intelligent algorithms, the present invention provides a high-precision and efficient intelligent detection method for the machining quality of CNC machine tool workpieces. By collecting the point cloud data of the workpiece to be detected and the standard CNC machine tool workpiece, and using the improved mean shift algorithm for clustering analysis, the abnormal regions can be accurately screened out, and potential defects can be carefully identified. The application of the ICP algorithm ensures the high-precision splicing of the point cloud data models of the workpiece to be detected and the standard workpiece. By calculating the area of the abnormal region through Boolean operations and combining it with the area threshold for judgment, the defect detection process becomes more accurate and reliable. In addition, preprocessing and optimization measures such as mean filtering and secondary clustering further improve the data quality and clustering accuracy, and enhance the stability and sensitivity in the quality detection of complex workpieces. This solution realizes the automatic detection and intelligent evaluation of the machining quality of CNC machine tool workpieces, significantly improving the detection efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By referring to the detailed description below with reference to the drawings, the above and other objects, features, and advantages of the exemplary embodiments of the present invention will become readily understood. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, where: Figure 1 is a flowchart schematically showing an intelligent detection method for the machining quality of CNC machine tool workpieces according to an embodiment of the present invention; Figure 2 is a structural block diagram schematically showing an intelligent detection system for the machining quality of CNC machine tool workpieces according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0022] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0023] The specific embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings.

[0024] An embodiment of an intelligent detection method for the machining quality of workpieces on a numerically controlled machine tool.

[0025] As Figure 1 shown, a flowchart of an intelligent detection method for the machining quality of workpieces on a numerically controlled machine tool according to an embodiment of the present invention includes the following steps: S1: Collect the point cloud data of each position of the workpiece to be detected on the numerically controlled machine tool and the standard workpiece on the numerically controlled machine tool.

[0026] In one embodiment, first, a binocular camera or a structured light camera is used to accurately collect each position of the workpiece to be detected on the numerically controlled machine tool and the standard workpiece to obtain complete three-dimensional point cloud data. The binocular camera can effectively capture the depth information of the surface details of the workpiece through the parallax difference between the two cameras and generate high-precision three-dimensional point clouds. The structured light camera can also obtain accurate three-dimensional surface data by projecting light with a known pattern and combining the deformation of the reflected light captured by the camera. Both of these technologies have high precision and stability in the process of point cloud data collection. Especially in the detection of complex shapes and precision machining surfaces, they can provide accurate depth information, which helps to improve the reliability of subsequent analysis results.

[0027] For the large amount of collected point cloud data, first perform mean filtering to remove abnormal points introduced by noise or unstable factors. Mean filtering can effectively reduce the interference of random noise by smoothing the data points, making the point cloud data smoother and more uniform, which helps to improve the precision of subsequent point cloud processing. In practical applications, point cloud data is often affected by factors such as environmental light changes, camera perspectives, and workpiece surface reflections, resulting in errors or noise in some point cloud data. Therefore, mean filtering can effectively avoid the influence of these noises on the point cloud data model, thereby providing a more accurate and clear basis for subsequent point cloud comparison and anomaly detection.

[0028] Furthermore, the point cloud data after mean filtering processing will have higher quality, which can provide reliable data support for the next steps of point cloud comparison, matching, and anomaly detection. When comparing with the point cloud data of the standard workpiece, the differences and anomalies on the surface of the workpiece to be detected can be effectively identified. These differences often manifest as deviations in size, shape, and position, or as minor defects generated during the machining process. After denoising the point cloud data by mean filtering, its surface contour and details are clearer, which can more accurately reflect the true shape of the workpiece, thereby improving the sensitivity and accuracy of anomaly detection.

[0029] S2: Use the improved mean shift algorithm to cluster the point cloud data of the workpiece to be detected, and remove the clusters that do not meet the DB index threshold to obtain the final point cloud data.

[0030] In one embodiment, the improved mean shift algorithm includes a drift vector, calculating the drift vector of the th data point belonging to the th cluster at the th iteration, , where is the initial drift vector of the th data point belonging to the th cluster at the th iteration, is the DB index of the th data point belonging to the th cluster at the th iteration, is the DB index of the th data point belonging to the th cluster at the th iteration, is an empirical constant.

[0031] An improved mean shift algorithm introduces a dynamic adjustment mechanism for the drift vector. Specifically, by comparing the DB indices of the clustering clusters at different iteration times, the drift vector is adjusted based on the change in the attribution of data points. This mechanism can capture the subtle changes in the data point clustering process more precisely, thereby improving the accuracy and stability of the clustering results. When the DB index of the data points changes between the old and new iterations, the drift vector will make corresponding adjustments according to the index difference, enabling the clustering clusters to more flexibly adapt to the distribution changes of the data points during the iteration process. By introducing the empirical constant G, the amplitude and speed of the drift vector adjustment are further optimized, avoiding the problems of overfitting or slow convergence in complex data sets. The improved algorithm can effectively improve the clustering accuracy. Especially when dealing with noisy data or complex data structures, it can achieve more accurate anomaly detection and data pattern recognition, thus significantly enhancing the performance and reliability of the model in practical applications.

[0032] In this embodiment, first, the improved mean shift algorithm is applied to cluster the point cloud data collected at various positions of the CNC machine tool workpiece to be detected, obtaining multiple clustering clusters. The improvement of the mean shift algorithm can automatically identify different data clusters according to the distribution of data points in the multi-dimensional space. Through continuous iteration, the algorithm can adjust the clustering center according to the density distribution of data points, so that the finally obtained clustering clusters can not only accurately reflect the overall structure of the data, but also effectively distinguish different regional features. Especially when there are small defects on the surface of complex workpieces, it can ensure the accurate identification of abnormal areas.

[0033] Next, for each clustering cluster, the clustering effect is evaluated by calculating its DB index. The DB index is an important indicator to measure the internal consistency of the clustering cluster and the separation degree between clusters. The smaller its value, the higher the compactness of the points within the cluster, the farther the distance between clusters, and the better the clustering effect. Specifically, the calculation process of the DB index includes two parts: First, the cluster center of each clustering cluster is determined by the mean method. The selection of the cluster center is based on the average position of the points within the cluster. This method has strong stability and low computational complexity. Second, the inter-cluster distance and intra-cluster distance of each clustering cluster are calculated. The inter-cluster distance refers to the distance between clustering clusters, while the intra-cluster distance reflects the tightness between points within the cluster. By calculating the mean of the inter-cluster distance and intra-cluster distance, the DB index of each clustering cluster is finally obtained.

[0034] Once the DB indices of all clustering clusters are obtained, the clusters are screened according to a preset threshold. If the DB index of a certain clustering cluster is less than the first set threshold, it indicates that the distribution of data points within this clustering cluster is relatively sparse and may belong to noise or invalid clusters. Therefore, this cluster will be removed to obtain more accurate point cloud data. In this way, irrelevant or meaningless clusters can be effectively filtered out, reducing the interference of noise on subsequent processing, thereby improving the efficiency and accuracy of data processing.

[0035] For those clusters with the DB index between the second set threshold and the first set threshold, secondary clustering will be performed. The purpose of secondary clustering is to further refine the distribution of data points in these clusters and avoid misclassifying some minor outliers as normal points. By performing meticulous secondary clustering on these clusters, more potential minor anomalies can be identified, improving the accuracy of anomaly detection. Secondary clustering can refine the classification structure of the point cloud, effectively distinguish some potential and imperceptible local defects or deformations, ensure the accuracy of the final obtained point cloud data, and thus enhance the reliability of detection.

[0036] S3: Stitch the point cloud data of the final workpiece to be detected and the point cloud data of each position of the standard CNC machine tool workpiece to obtain the point cloud data model of the workpiece to be detected and the standard CNC machine tool workpiece.

[0037] In one embodiment, the ICP algorithm optimizes the point cloud registration result by repeatedly iterating and utilizing the geometric similarity between the workpiece to be detected and the standard workpiece, and finally achieves the accurate alignment of the two sets of point cloud data.

[0038] Specifically, the core idea of the ICP algorithm is to continuously find the corresponding nearest points between the point cloud to be aligned and the target point cloud during the iteration process, and calculate the transformation matrix that minimizes the error based on these matching points, thereby achieving the spatial alignment of the two point clouds. The algorithm gradually adjusts the position and orientation of the source point cloud by calculating the distance difference between the nearest point pairs of the source point cloud and the target point cloud, and finally makes the overlapping area of the two point clouds as close as possible. This method based on minimum error optimization can effectively eliminate the initial pose error and improve the registration accuracy. Especially when there is a certain deviation in the initial position of the point cloud data, the ICP algorithm can gradually converge to the optimal solution through multiple iterations, ensuring that the final stitched point cloud data model has high accuracy.

[0039] After applying the ICP algorithm, the point cloud data of the workpiece to be detected and the standard CNC machine tool workpiece will be aligned and fused into a complete point cloud data model. Through this stitching method, seamless docking of the two workpieces in the same coordinate system can be achieved, providing a reliable three-dimensional data basis for subsequent shape comparison, defect detection, and quality assessment. This precise registration of the point cloud data not only improves the accuracy of the workpiece model but also provides a high-resolution three-dimensional view for the detection of minor defects, ensuring higher accuracy and sensitivity in judging the surface quality of the workpiece.

[0040] Through the splicing process of the ICP algorithm, it is possible to effectively eliminate the deviations or displacements that may occur during the measurement of the workpiece, reduce the errors caused by data noise, sensor errors, or different measurement angles, thereby improving the overall quality of the point cloud data model. In addition, the adaptability of the ICP algorithm enables it to operate stably under different initial conditions and point cloud data qualities, ensuring the robustness of the registration process. This not only makes the technical solution more flexible and reliable in practical applications but also provides consistent performance in different industrial environments.

[0041] S4: Perform a Boolean operation on the point cloud data model to obtain the area of the abnormal region. If the area of the abnormal region is greater than the set area threshold, it is determined that the CNC machine tool workpiece to be detected is abnormal.

[0042] The value of the above-mentioned set area threshold can be 0.5. Of course, it can also be determined according to the actual situation.

[0043] In one embodiment, these two point cloud data models are processed through Boolean operations to identify the abnormal regions therein. Boolean operations, as a common geometric processing method, can efficiently perform operations such as intersection, union, and difference on two three-dimensional models, thereby finding the geometric differences and abnormal regions between the workpiece to be detected and the standard workpiece.

[0044] In this process, the Boolean operation will compare the two based on the geometric structures between the point cloud models and calculate the difference regions between them. Specifically, the Boolean operation can effectively highlight the differences between the workpiece to be detected and the standard workpiece in terms of shape, size, position, etc. Especially when there are wear, defects, or deformations on the surface of the workpiece, the Boolean operation can clearly mark these abnormal regions. These difference regions usually manifest as some subtle geometric deformations, such as local depressions, protrusions, or cracks, etc., and the Boolean operation can clearly extract the abnormal regions by removing the overlapping parts of these regions with the standard model.

[0045] Once these abnormal regions are obtained, their impact degree is evaluated by calculating the area of these regions. If the area of the abnormal region is greater than the preset area threshold, it means that the abnormality in this region is sufficient to affect the quality and function of the workpiece, and then it is determined that the workpiece to be detected is an abnormal workpiece. The setting basis of the area threshold can be the design standard of the workpiece or the tolerance range specified by the industry regulations, ensuring that only those abnormal regions that exceed the normal tolerance range are regarded as significant defects. This judgment process has high sensitivity and can accurately identify those subtle but important defect regions, thereby ensuring that the detection can provide consistent and accurate judgment results under various workpiece conditions.

[0046] By combining Boolean operations with area threshold determination, the quality of the workpiece of the numerically controlled machine tool to be detected can be comprehensively and accurately evaluated. In practical applications, the high efficiency and calculation accuracy of Boolean operations can process large-scale point cloud data while ensuring the reliability of the detection results. By introducing the area threshold, false positives and false negatives can be avoided during anomaly detection, ensuring that only the abnormal areas that truly affect the function and performance of the workpiece are marked as defective areas, thus improving the accuracy and practicality of the entire detection process.

[0047] The solution of the present invention realizes accurate workpiece quality detection through efficient point cloud data acquisition, processing and analysis. By using an improved mean shift algorithm to cluster the point cloud data, abnormal areas can be effectively screened out. Combining with the ICP algorithm to accurately register the point cloud data models of the workpiece to be detected and the standard workpiece, and then calculating the area of the abnormal area through Boolean operations to ensure accurate identification of defects. In addition, secondary clustering further improves the accuracy of anomaly detection, and the application of mean filtering makes the point cloud data smoother and more reliable. The overall solution can automatically detect the micro-defects of the workpiece, improving the accuracy, efficiency and stability of the detection, and meeting the requirements of high-precision numerical control machine tool processing quality monitoring.

[0048] An embodiment of an intelligent detection system for the machining quality of numerically controlled machine tool workpieces: As Figure 2 shown, the structural block diagram of an intelligent detection system for the machining quality of numerically controlled machine tool workpieces according to an embodiment of the present invention includes a processor and a memory.

[0049] The present invention also provides an intelligent detection system for the machining quality of numerically controlled machine tool workpieces. As Figure 2 shown, the system includes a processor and a memory, and the memory stores computer program instructions, which, when executed by the processor, implement an intelligent detection method for the machining quality of numerically controlled machine tool workpieces according to the above of the present invention.

[0050] The intelligent detection system for the machining quality of numerically controlled machine tool workpieces further includes other components well known to those skilled in the art such as a communication interface, and its settings and functions are known in the art, so they will not be elaborated here.

[0051] In the present invention, the foregoing memory may be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory (RRAM), a dynamic random access memory (DRAM), a static random access memory (SRAM), an enhanced dynamic random access memory (EDRAM), a high-bandwidth memory (HBM), a hybrid memory cube (HMC), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions stored or otherwise maintained by such a computer-readable medium.

[0052] In the description of this specification, the meanings of "a plurality of" and "several" are at least two, such as two, three, or more, etc., unless otherwise specifically defined.

[0053] Although this specification has shown and described multiple embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in the practice of the present invention.

Claims

1. An intelligent detection method for machining quality of CNC machine tools, characterized in that: include: Collect point cloud data of each position of the CNC machine tool workpiece to be inspected and the standard CNC machine tool workpiece; The point cloud data of each position of the CNC machine tool workpiece to be detected are clustered by using an improved mean shift algorithm to obtain a plurality of clusters; the DB index of each cluster is obtained, and if the DB index is less than a first set threshold, the corresponding cluster is eliminated to obtain the final point cloud data; the final point cloud data and the point cloud data of each position of the standard CNC machine tool workpiece are spliced ​​to obtain the point cloud data model of the CNC machine tool workpiece to be detected and the standard CNC machine tool workpiece; Performing Boolean operations on the point cloud data models of the CNC machine tool workpiece to be detected and the standard CNC machine tool workpiece to obtain the area of ​​the abnormal region. If the area of ​​the abnormal region is greater than a set area threshold, the CNC machine tool workpiece to be detected is determined to be abnormal. The improved mean shift algorithm includes a drift vector, which is calculated in the first The first iteration The data point belongs to The drift vector of the clusters , , where For the The first iteration The data point belongs to The initial drift vector of the clusters, For the The first iteration The data point belongs to The DB index of the clusters, For the The first iteration The data point belongs to The DB index of the clusters, is an empirical constant.

2. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: The DB index of each cluster is obtained as follows: The mean method is used to obtain the cluster center of each cluster, and the inter-cluster distance and intra-cluster distance of each cluster are calculated; The mean of the inter-cluster distance and intra-cluster distance of each cluster is taken as the DB index of the cluster.

3. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: A binocular camera or a structured light camera is used to collect point cloud data of each position of the CNC machine tool workpiece to be inspected and the standard CNC machine tool workpiece.

4. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: The method also includes performing mean filtering on the point cloud data.

5. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: The final point cloud data and the point cloud data of each position of the standard CNC machine tool workpiece are spliced ​​using the ICP algorithm.

6. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: The set area threshold is 0.

5.

7. The method for intelligent detection of machining quality of workpieces of CNC machine tools according to claim 1, characterized in that: The method further includes performing secondary clustering on the point cloud data in the corresponding clustering cluster whose DB index is greater than the second set threshold and less than the first set threshold.

8. An intelligent detection system for machining quality of CNC machine tools, characterized in that: The method comprises a memory and a processor, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an intelligent detection method for machining quality of a CNC machine tool workpiece as claimed in any one of claims 1 to 7 is implemented.

Images