A machine tool tool recognition and analysis method based on visual analysis

Through visual analysis technology, real-time identification and verification of machine tool tools has been solved, and the losses and risks caused by error identification in the prior art are improved, and the identification accuracy and safety are improved.

CN118438264BActive Publication Date: 2024-11-05SHENZHEN LINGZHIGUANG MECHANICAL & ELECTRICAL AUTOMATION SYST CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410622880.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-05-20
Publication Date
2024-11-05
Estimated Expiration
2044-05-20

AI Technical Summary

Technical Problem

Existing machine tool tool recognition and analysis methods are usually only used before use, which can easily lead to identification errors, resulting in batch damage to the tool or processed products, and even damage to the machine tool.

Method used

The machine tool tool recognition and analysis method based on visual analysis is adopted to identify and verify tools in real time through image acquisition, preprocessing, feature extraction, verification and manual verification to ensure the accuracy of recognition.

Benefits of technology

The accuracy of machine tool recognition is improved, losses and risks caused by identification errors are avoided, and the accuracy of identification is guaranteed through multiple verifications.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118438264B_ABST
    Figure CN118438264B_ABST
Patent Text Reader

Abstract

The present invention discloses a machine tool tool identification and analysis method based on visual analysis, which relates to the technical field of machine tool tool identification and is used to solve the problem that the existing machine tool tool identification and analysis method usually only identifies the tool before use, and once the identification is wrong, it will lead to greater losses and risks, such as batch damage of the tool or processed products, and in severe cases, it will also cause greater damage to the machine tool. The present invention comprises the following steps: S1: firstly, a tool image of the working area of ​​the machine tool is obtained, and the obtained tool image is preprocessed; the present invention performs different identification inspections on the machine tool tool at different time periods, namely, visual identification after the tool is installed, tool path planning identification before the machine tool is started, and product feature identification after the product is installed. The tool identification and verification method of the three time periods can greatly improve the accuracy of machine tool tool identification, and can also avoid greater losses caused by wrong tools.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of machine tool tool identification, and in particular to a machine tool tool identification and analysis method based on visual analysis. Background Art

[0002] In industrial production, machine tools are widely used in various fields. The tools they use are of various types and similar, and accurate tool identification is required to ensure the quality and efficiency of processing. Traditional tool identification methods mainly rely on manual detection, which has problems such as complex operation, slow speed, and easy errors.

[0003] The existing machine tool tool recognition and analysis methods usually only identify the tool before use. Once the recognition is wrong, it will lead to greater losses and risks, such as batch damage of the tool or processed products. In severe cases, it will cause greater damage to the machine tool. Therefore, a machine tool tool recognition and analysis method based on visual analysis is designed.

[0004] In order to solve the above defects, a technical solution is now provided. Summary of the invention

[0005] The purpose of the present invention is to solve the problem that the existing machine tool tool identification and analysis methods usually only identify the tool before use. Once the identification is wrong, it will cause great losses and risks, such as batch damage of the tool or processed products, and in severe cases, it will cause great damage to the machine tool. A machine tool tool identification and analysis method based on visual analysis is proposed.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A machine tool tool recognition and analysis method based on visual analysis includes the following steps:

[0008] S1: firstly, obtaining a tool image in the working area of ​​the machine tool and preprocessing the obtained tool image;

[0009] S2: extracting feature information from the tool image using computer vision methods, including the shape, texture and identification features of the tool;

[0010] S3: then verify the extracted tool image features respectively to obtain the tool image data stored in the database with the highest similarity, and store the tool information inconsistent with the tool image data in the database as new tool data in the database;

[0011] S4: Verify the identified tool to be used, analyze and determine whether the tool identification is correct according to the tool path planning, and start controlling the machine tool to control the tool to process the product according to the real-time tool path planning if the identification is correct. If there is an abnormality in the identification, perform manual verification operation;

[0012] S5: Finally, the tools with abnormalities are manually checked through the manual check platform.

[0013] Further, a tool identification system is included, and the tool identification system includes:

[0014] An image acquisition module is used to acquire tool images in the working area of ​​a machine tool through a camera or other image acquisition device, and to preprocess the acquired tool images; the preprocessing specifically includes denoising, enhancement, and edge detection operations;

[0015] The feature extraction module is used to extract feature information from the tool image using computer vision methods, and extract the shape, texture and identification features of the tool by using SIFT and HOG feature extraction algorithms;

[0016] The recognition module is used to classify and identify the extracted features, verify the shape, texture and identification features of the tool, and identify the type of tool based on the tool detection in the database;

[0017] The verification module is used to verify the identified tool to be used, determine whether the identified tool is correct, and generate a manual verification signal if there is an abnormality;

[0018] The manual verification module is used to manually and remotely identify tools with identification anomalies by auditors on the manual verification platform.

[0019] Furthermore, the specific operation steps of the recognition module to verify the shape and texture of the tool image are as follows:

[0020] The verification steps for the tool image shape are as follows:

[0021] First, the tool image shape is segmented into three parts: the tool handle, the blade and the tool tip. The shapes of the segmented parts of the tool are quantified. Specifically, the shape size, angle and shape curvature of the segmented parts of the tool are obtained and calibrated as XC, JD and XQ respectively. After normalization, they are entered into the following formula: XBZ=1.23(XC+JD) 1.11 +XQ-0.22, to obtain the shape mark value XBZ of the tool segmentation part;

[0022] The shape mark values ​​of the three parts of the tool are XBZ1, XBZ2 and XBZ3 respectively, which are normalized and then substituted into the following formula: To get the total value ZBZ, where are preset weight coefficients for the tool handle shape value, the blade shape value, and the tool tip shape value, and are 0.993, 1.224, and 1.125, respectively;

[0023] Then the obtained shape total value ZBZ is compared with the tool image data and the corresponding shape total value pre-entered in the database, and the tool image data corresponding to the three most similar shape total values ​​are obtained, and the three tool image data are marked, and texture verification is performed on the marked three tool image data;

[0024] The verification steps for tool image texture are as follows:

[0025] Firstly, the blade image data is intercepted from the tool image data, and the texture analysis is performed on the blade image data. The blade image data is segmented again and divided into several small areas. The distribution of the pixel gray value in each area is calculated, so as to obtain the gray histogram of the blade surface in the entire blade image data.

[0026] Then, the local binary pattern algorithm (LBP) is used to extract the texture features of the blade image data. By comparing the 8 pixels around the pixel with the central pixel, a binary number is generated to represent the texture type of the pixel. The LBP feature extraction is performed on the entire image to obtain a set of feature vectors representing the blade texture.

[0027] The obtained blade surface grayscale histogram and a set of blade texture feature vectors are respectively compared with the blade surface grayscale histogram and a set of blade texture feature vectors corresponding to three tool image data marked in the database to obtain a tool image data with the highest similarity;

[0028] The tool image data with the highest similarity is determined to be the identified tool, and the information of the tool in the database is obtained.

[0029] Furthermore, the specific operation steps of the identification module to verify the tool identification information are as follows:

[0030] When identification information is detected in the tool image data, the tool identification information in the tool image data is obtained, wherein the tool identification information includes: tool model, production date, manufacturer or serial number, and the tool identification information obtained in real time is checked with the corresponding identification information in the tool image data with the highest similarity through the identification edge detection method to determine whether they are consistent. When the tool information is consistent, the verification is completed, and it is determined that the tool image data with the highest similarity is the identified tool;

[0031] When the tool information is inconsistent, the identified tool information is determined to be an unrecorded tool, and the unrecorded tool is taken as a new tool and the tool image data and identification information are obtained and stored in a database.

[0032] Furthermore, the specific operation steps of the verification module for verifying the identified tool to be used are as follows:

[0033] After the tool is identified, all tool path plans of the tool in the database are obtained. The tool path plan specifically includes the motion trajectory, speed, acceleration and angle of the machine tool control tool. Then, the real-time tool path plan before the machine tool control tool is started for processing is compared. The real-time tool path plan is compared with all tool path plans of the tool stored in the database. When all tool path plans of the tool stored in the database are the same as the real-time tool path plan, the tool is verified, and it is determined that the tool is correctly identified and the operation command is executed through the machine tool.

[0034] When all tool path plans for the tool stored in the database do not have the same tool path plan as the real-time tool path plan, the identification of the tool is further verified. The specific steps are as follows:

[0035] Get the real-time tool path planning of this tool, and obtain the information of the processed product. The specific processing product information includes product material, size and geometry, surface flatness and surface roughness. Calculate the volume and surface area of ​​the processed product through the size and geometry in the processing product information, and normalize them together with the surface flatness and surface roughness of the processed product, and substitute them into the following formula: , to obtain the quality value dxz, where tj, bj, pz and cc are the volume, surface area, surface flatness and surface roughness of the processed product respectively;

[0036] The obtained quality value dxz is respectively calculated with the quality value of different processed products in the database, and the absolute value of each difference is processed to obtain the absolute difference value. The obtained absolute difference values ​​are compared with the preset difference threshold. When there is an absolute difference value less than the preset difference threshold, the corresponding processed product is marked, and the marked processed product is further determined. The specific steps are as follows:

[0037] Compare the product material in the acquired processing product information with the product material corresponding to the marked processing product. When the same product material exists, obtain the processing history data of the product, and compare the tool path planning corresponding to the processing of the product with the real-time tool path planning. When the tool path planning is consistent, it is determined that the tool identification is correct, and start controlling the machine tool to control the tool to process the product according to the real-time tool path planning;

[0038] When there is no absolute difference value less than the preset difference threshold, a manual verification signaling is generated and sent to the manual verification module.

[0039] Furthermore, the manual verification module performs manual remote identification of tools with identification anomalies through the auditors of the manual verification platform in the following specific operation steps:

[0040] The tool image data with the highest similarity determined in the identified tool information and the real-time tool image data are sent to the manual verification platform. When collecting the real-time tool image, the tool is controlled to rotate slowly through the machine tool, and the slowly rotating tool is recorded in multiple directions. The tool image data with the highest similarity is then integrated with the real-time tool image data through the manual verification platform, and the same parts in the two sets of images are moved to the same display area and the image names are calibrated respectively.

[0041] The work status of the current auditors is then analyzed through the manual verification platform. If there are auditors who are idle, the calibrated image data and multi-directional video data of the tool are sent to the auditors’ terminals together with the inspection receipt.

[0042] When all auditors are in the auditing state, the estimated remaining completion time of all auditors is analyzed, and all the estimated remaining completion time obtained is compared with the preset waiting time threshold. When there is an estimated remaining completion time lower than the preset waiting time threshold, the corresponding auditor is marked, and the audit data is sent to the marked auditor's terminal for review;

[0043] When there is no estimated remaining completion time lower than the preset waiting time threshold, the urgency of the product processed by the machine tool is evaluated, and the specific steps are as follows:

[0044] Collect the remaining delivery period, number of products to be processed, single-piece processing time, processing difficulty and machine idle time of the product. The processing difficulty of the product is determined by analyzing the processing technology and material type of the product, and the processing difficulty is divided into five levels, and the five levels correspond to five degree values ​​1, 2, 3, 4 and 5 respectively;

[0045] Then, the remaining delivery period, the number of products to be processed, the processing time of a single piece, the corresponding degree values ​​of the processing difficulty and the idle time of the machine tool are calibrated as jh, js, sc, dz and xz respectively, and after normalization, they are entered into the following formula: To obtain the emergency value JXZ, and compare the obtained product emergency value with the tools that all auditors are reviewing and the corresponding product preset emergency values, obtain the preset emergency value lower than the product emergency value JXZ and the corresponding auditor and mark them, calculate the difference between the product emergency value JXZ and the marked preset emergency value, compare the obtained several differences to get the largest difference and the corresponding auditor, then send the audit data to the auditor's terminal, prompt the auditor to give priority to reviewing the tool corresponding to the product, and obtain the auditor's inspection receipt to determine whether the tool identification is correct.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] (1) The present invention performs different identification inspections on machine tool tools at different time periods, namely, visual identification after tool installation, tool path planning identification before machine tool startup, and product feature identification after product installation. The tool identification and verification methods at three time periods can greatly improve the accuracy of machine tool tool identification and avoid large losses caused by incorrect tools.

[0048] (2) In the present invention, when an abnormality occurs during the tool identification process, the tool with the abnormal identification can be re-examined through manual review in the manual verification platform, thereby providing multiple guarantees to ensure the accuracy of machine tool tool identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0050] Figure 1 is a flow chart of the method of the present invention;

[0051] Figure 2 This is the overall block diagram of the tool identification system in the present invention. DETAILED DESCRIPTION

[0052] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] It should be understood that the terms "include" and "comprising" used in the specification and claims of the present disclosure indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or collections thereof.

[0054] It should also be understood that the terms used in this disclosure are only for the purpose of describing specific embodiments and are not intended to limit the disclosure. As used in this disclosure and claims, the singular forms of "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise. It should also be further understood that the term "and / or" used in this disclosure and claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0055] like Figure 1-Figure 2As shown, a machine tool tool recognition and analysis method based on visual analysis includes a tool recognition system, which is composed of an image acquisition module, a feature extraction module, a recognition module, a verification module and a manual verification module;

[0056] The image acquisition module is used to acquire the tool image in the working area of ​​the machine tool and pre-process the acquired tool image. A camera or other image acquisition device can be used for real-time acquisition.

[0057] The acquired tool image is then preprocessed, including denoising, enhancement, and edge detection operations, to improve the accuracy of subsequent feature extraction and classification recognition.

[0058] The feature extraction module is used to use computer vision methods to extract feature information from tool images, including: the shape, texture and identification features of the tool. Traditional feature extraction algorithms such as SIFT or HOG can be used. SIFT is a scale-invariant feature extraction algorithm suitable for object recognition and image matching under different scales, rotations and lighting conditions. The SIFT algorithm detects local extreme points and extracts descriptors around these points to represent features. These descriptors have good robustness to rotation, scale and partial occlusion, and are widely used in tasks such as image feature matching and object recognition. HOG is a feature descriptor based on local gradient direction. The HOG algorithm divides the image into small local areas, counts the gradient direction histograms in each area, and concatenates these histograms as the final feature vector. The HOG feature is mainly used in tasks such as target detection and pedestrian recognition, and can effectively describe the edge and texture features of objects. Deep learning methods can also be used for feature extraction.

[0059] The recognition module is used to classify and identify the extracted features and identify the type of tool;

[0060] Verify the tool image shape and texture. The verification steps for the tool image shape are as follows:

[0061] First, the tool image shape is segmented into three parts: the handle, the blade and the tip. The shapes of the segmented parts of the tool are quantified. Specifically, the shape size, angle and shape curvature of the segmented parts of the tool are obtained and calibrated as XC, JD and XQ respectively. After normalization, they are entered into the following formula: XBZ=1.23(XC+JD) 1.11 +XQ-0.22 to get the shape value XBZ of the tool segmentation part, and the shape values ​​of the three segmentation parts of the tool are XBZ1, XBZ2 and XBZ3 respectively. After normalization, substitute them into the following formula: To get the total value ZBZ, where are respectively the preset weight coefficients of the tool handle shape value, the preset weight coefficients of the blade shape value and the preset weight coefficients of the tool tip shape value, and the values ​​are 0.993, 1.224 and 1.125 respectively; and the shape total value is used as the standard for quantifying the tool shape; the obtained shape total value ZBZ is compared with the tool image data and the corresponding shape total value pre-entered in the database, and the tool image data corresponding to the three closest shape total values ​​are obtained, and the three tool image data are marked, and texture verification is performed on the marked three tool image data;

[0062] The verification steps for tool image texture are as follows:

[0063] Firstly, the blade image data is intercepted from the tool image data, and the texture analysis is performed on the blade image data. The blade image data is segmented again and divided into several small areas, and the distribution of the pixel grayscale value in each area is calculated, so as to obtain the grayscale histogram of the blade surface in the entire blade image data; then the local binary pattern algorithm, i.e., LBP, is used to extract the texture features of the blade image data, and a binary number is generated by comparing the 8 pixels around the pixel point with the central pixel, so as to represent the texture type of the pixel point, and the LBP feature extraction is performed on the entire image to obtain a set of feature vectors representing the blade texture; the obtained blade surface grayscale histogram and a set of blade texture feature vectors are respectively checked with the blade surface grayscale histogram and a set of blade texture feature vectors corresponding to the three tool image data marked in the database, so as to obtain a tool image data with the highest similarity; then the tool image data with the highest similarity is determined to be the identified tool, and the information of the tool in the database is obtained;

[0064] When identification information is detected in the tool image data, the tool identification information is verified again. The specific steps are as follows: based on the tool identification information in the tool image data obtained, the tool identification information includes: tool model, production date, manufacturer or serial number, and the tool identification information obtained in real time is checked with the corresponding identification information in the tool image data with the highest similarity through the identification edge detection method to determine whether they are consistent. When the tool information is consistent, the verification is completed and the tool image data with the highest similarity is determined to be the identified tool; when the tool information is inconsistent, the identified tool information is determined to be an unentered tool, and the unentered tool is taken as a new tool and the tool image data and identification information are obtained and stored in the database.

[0065] A verification module, used to verify the identified tool to be used;

[0066] After completing the identification of the tool, obtain all the tool path plans of the tool in the database. The tool path plan specifically includes the motion trajectory, speed, acceleration and angle of the machine tool control tool. Then compare the real-time tool path plan before starting the machine tool control tool for processing. Compare the real-time tool path plan with all the tool path plans of the tool stored in the database. When all the tool path plans of the tool stored in the database are the same as the real-time tool path plan, the verification of the tool is completed, and it is determined that the tool identification is correct and the operation command is executed through the machine tool. When all the tool path plans of the tool stored in the database are not the same as the real-time tool path plan, the identification of the tool is further verified. The specific steps are as follows:

[0067] Get the real-time tool path planning of this tool, and obtain the information of the processed product. The specific processing product information includes product material, size and geometry, surface flatness and surface roughness. Calculate the volume and surface area of ​​the processed product through the size and geometry in the processing product information, and normalize them together with the surface flatness and surface roughness of the processed product, and substitute them into the following formula: , to obtain the quality signal value dxz, where tj, bj, pz and cc are the volume, surface area, surface flatness and surface roughness of the processed product respectively; the obtained quality signal value dxz is respectively calculated with the quality signal values ​​of different processed products in the database, and each difference is processed to obtain the absolute value of the difference, and the obtained absolute values ​​of the difference are compared with the preset difference threshold. When there is an absolute value of the difference less than the preset difference threshold, the corresponding processed product is marked, and the marked processed product is further determined. The specific steps are as follows: the product material in the obtained processed product information is compared with the product material corresponding to the marked processed product. When the same product material exists, the processing history data of the product is obtained, and the tool path planning corresponding to the processing of the product is compared with the real-time tool path planning. When the tool path planning is consistent, it is judged that the tool identification is correct, and the control machine tool is started to control the tool to process the product according to the real-time tool path planning;

[0068] When there is no absolute difference value less than the preset difference threshold, a manual verification signaling is generated and sent to the manual verification module.

[0069] A manual verification module is used to receive the manual verification signal generated by the inspection module, and perform manual verification operations through the manual verification signal, and the auditors of the manual verification platform manually remotely identify the tools with recognition abnormalities;

[0070] The tool image data with the highest similarity and the real-time tool image data in the identified tool information are sent to the manual verification platform. When collecting the real-time tool image, the tool is controlled to rotate slowly through the machine tool, and the slowly rotating tool is recorded in multiple directions. The tool image data with the highest similarity and the real-time tool image data are then integrated through the manual verification platform, and the same parts in the two sets of images are moved to the same display area and the image names are calibrated respectively, so as to facilitate the distinction of the auditors;

[0071] The work status of the current auditors is then analyzed through the manual verification platform. When there are auditors in an idle state, the calibrated image data and multi-directional video data of the tool are sent to the idle auditor's terminal together with the inspection receipt; when all auditors are in the audit state, the estimated remaining completion time of all auditors is analyzed, and all the estimated remaining completion time obtained is compared with the preset waiting time threshold. When there is an estimated remaining completion time lower than the preset waiting time threshold, the corresponding auditor is marked, and the audit data is sent to the marked auditor's terminal for review. When there is no estimated remaining completion time lower than the preset waiting time threshold, When the time is short, the urgency of the product processed by the machine tool is evaluated. The specific steps are as follows: collect the remaining delivery period of the product, the number of products to be processed, the processing time of a single piece, the processing difficulty and the idle time of the machine tool. The product processing difficulty is determined by analyzing the processing technology and material type of the product, and the processing difficulty is divided into five levels. The higher the level, the greater the corresponding processing difficulty. The five levels correspond to five degree values ​​1, 2, 3, 4 and 5 respectively; then the remaining delivery period, the number of products to be processed, the processing time of a single piece, the degree values ​​corresponding to the processing difficulty and the idle time of the machine tool are calibrated as jh, js, sc, dz and xz respectively, and after normalization, they are entered into the following formula: To obtain the emergency value JXZ, and compare the obtained product emergency value with the tools that all auditors are reviewing and the corresponding product preset emergency values, obtain the preset emergency value lower than the product emergency value JXZ and the corresponding auditor and mark them, calculate the difference between the product emergency value JXZ and the marked preset emergency value, compare the obtained several differences to get the largest difference and the corresponding auditor, then send the audit data to the auditor's terminal, prompt the auditor to give priority to reviewing the tool corresponding to the product, and obtain the auditor's inspection receipt to determine whether the tool identification is correct.

[0072] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A machine tool tool recognition and analysis method based on visual analysis, characterized in that: The following steps are involved: S1: firstly, obtaining a tool image in the working area of ​​the machine tool and preprocessing the obtained tool image; S2: extracting feature information from the tool image using computer vision methods, including the shape, texture and identification features of the tool; S3: then verify the extracted tool image features respectively to obtain the tool image data stored in the database with the highest similarity, and store the tool information inconsistent with the tool image data in the database as new tool data in the database; S4: Verify the identified tool to be used, analyze and determine whether the tool identification is correct according to the tool path planning, and start controlling the machine tool to control the tool to process the product according to the real-time tool path planning if the identification is correct. If there is an abnormality in the identification, perform manual verification operation; S5: Finally, the tool with abnormality is manually checked through the manual checking platform; Including tool identification system, tool identification system includes: An image acquisition module is used to acquire tool images in the working area of ​​a machine tool through a camera or other image acquisition device, and to preprocess the acquired tool images; the preprocessing specifically includes denoising, enhancement, and edge detection operations; The feature extraction module is used to extract feature information from the tool image using computer vision methods, and extract the shape, texture and identification features of the tool by using SIFT and HOG feature extraction algorithms; The recognition module is used to classify and identify the extracted features, verify the shape, texture and identification features of the tool, and identify the type of tool based on the tool detection in the database; The verification module is used to verify the identified tool to be used, determine whether the identified tool is correct, and generate a manual verification signal if there is an abnormality; The manual verification module is used to manually and remotely identify tools with abnormal identification by auditors on the manual verification platform; The specific operation steps of the recognition module to verify the shape and texture of the tool image are as follows: The verification steps for the tool image shape are as follows: First, the tool image shape is segmented into three parts: the tool handle, the blade and the tool tip. The shapes of the segmented parts of the tool are quantified. Specifically, the shape size, angle and shape curvature of the segmented parts of the tool are obtained and calibrated as XC, JD and XQ respectively. After normalization, they are entered into the following formula: XBZ=1.23(XC+JD) 1.11 +XQ-0.22, to obtain the shape mark value XBZ of the tool segmentation part; The shape mark values ​​of the three parts of the tool are XBZ1, XBZ2 and XBZ3 respectively, which are normalized and then substituted into the following formula: To get the total value ZBZ, where They are respectively the preset weight coefficients of the tool handle shape mark value, the blade shape mark value and the tool tip shape mark value; Then the obtained shape total value ZBZ is compared with the tool image data and the corresponding shape total value pre-entered in the database, and the tool image data corresponding to the three most similar shape total values ​​are obtained, and the three tool image data are marked, and texture verification is performed on the marked three tool image data; The verification steps for tool image texture are as follows: Firstly, the blade image data is intercepted from the tool image data, and the texture analysis is performed on the blade image data. The blade image data is segmented again and divided into several small areas. The distribution of the pixel gray value in each area is calculated, so as to obtain the gray histogram of the blade surface in the entire blade image data. Then, the local binary pattern algorithm (LBP) is used to extract the texture features of the blade image data. By comparing the 8 pixels around the pixel with the central pixel, a binary number is generated to represent the texture type of the pixel. The LBP feature extraction is performed on the entire image to obtain a set of feature vectors representing the blade texture. The obtained blade surface grayscale histogram and a set of blade texture feature vectors are respectively compared with the blade surface grayscale histogram and a set of blade texture feature vectors corresponding to three tool image data marked in the database to obtain a tool image data with the highest similarity; Then the tool image data with the highest similarity is determined to be the identified tool, and the information of the tool in the database is obtained; The specific operation steps of the verification module for verifying the identified tool to be used are as follows: After the tool is identified, all tool path plans of the tool in the database are obtained. The tool path plan specifically includes the motion trajectory, speed, acceleration and angle of the machine tool control tool. Then, the real-time tool path plan before the machine tool control tool is started for processing is compared. The real-time tool path plan is compared with all tool path plans of the tool stored in the database. When all tool path plans of the tool stored in the database are the same as the real-time tool path plan, the tool is verified, and it is determined that the tool is correctly identified and the operation command is executed through the machine tool. When all tool path plans for the tool stored in the database do not have the same tool path plan as the real-time tool path plan, the identification of the tool is further verified. The specific steps are as follows: Get the real-time tool path planning of this tool, and obtain the information of the processed product. The specific processing product information includes product material, size and geometry, surface flatness and surface roughness. Calculate the volume and surface area of ​​the processed product through the size and geometry in the processing product information, and normalize them together with the surface flatness and surface roughness of the processed product, and substitute them into the following formula: To obtain the quality value dxz, where tj, bj, pz and cc are the volume, surface area, surface flatness and surface roughness of the processed product respectively; The obtained quality value dxz is respectively calculated with the quality value of different processed products in the database, and the absolute value of each difference is processed to obtain the absolute difference value. The obtained absolute difference values ​​are compared with the preset difference threshold. When there is an absolute difference value less than the preset difference threshold, the corresponding processed product is marked, and the marked processed product is further determined. The specific steps are as follows: Compare the product material in the acquired processing product information with the product material corresponding to the marked processing product. When the same product material exists, obtain the processing history data of the product, and compare the tool path planning corresponding to the processing of the product with the real-time tool path planning. When the tool path planning is consistent, it is determined that the tool identification is correct, and start controlling the machine tool to control the tool to process the product according to the real-time tool path planning; When there is no absolute difference value less than the preset difference threshold, a manual verification signaling is generated and sent to the manual verification module.

2. The machine tool tool recognition and analysis method based on visual analysis according to claim 1, characterized in that: The specific operation steps of the recognition module to verify the tool identification information are as follows: When identification information is detected in the tool image data, the tool identification information in the tool image data is obtained, wherein the tool identification information includes: tool model, production date, manufacturer or serial number, and the tool identification information obtained in real time is checked with the corresponding identification information in the tool image data with the highest similarity through the identification edge detection method to determine whether they are consistent. When the tool information is consistent, the verification is completed, and it is determined that the tool image data with the highest similarity is the identified tool; When the tool information is inconsistent, the identified tool information is determined to be an unrecorded tool, and the unrecorded tool is taken as a new tool and the tool image data and identification information are obtained and stored in a database.

3. The machine tool tool recognition and analysis method based on visual analysis according to claim 1, characterized in that: The specific operation steps of the manual verification module for the reviewers of the manual verification platform to manually remotely identify the tools with recognition anomalies are as follows: The tool image data with the highest similarity determined in the identified tool information and the real-time tool image data are sent to the manual verification platform. When collecting the real-time tool image, the tool is controlled to rotate slowly through the machine tool, and the slowly rotating tool is recorded in multiple directions. The tool image data with the highest similarity is then integrated with the real-time tool image data through the manual verification platform, and the same parts in the two sets of images are moved to the same display area and the image names are calibrated respectively. The work status of the current auditors is then analyzed through the manual verification platform. If there are auditors who are idle, the calibrated image data and multi-directional video data of the tool are sent to the auditors’ terminals together with the inspection receipt. When all auditors are in the auditing state, the estimated remaining completion time of all auditors is analyzed, and all the estimated remaining completion time obtained is compared with the preset waiting time threshold. When there is an estimated remaining completion time lower than the preset waiting time threshold, the corresponding auditor is marked, and the audit data is sent to the marked auditor's terminal for review; When there is no estimated remaining completion time lower than the preset waiting time threshold, the urgency of the product processed by the machine tool is evaluated, and the specific steps are as follows: Collect the remaining delivery period, number of products to be processed, single-piece processing time, processing difficulty and machine idle time of the product. The processing difficulty of the product is determined by analyzing the processing technology and material type of the product, and the processing difficulty is divided into five levels, and the five levels correspond to five degree values ​​1, 2, 3, 4 and 5 respectively; Then, the remaining delivery period, the number of products to be processed, the processing time of a single piece, the corresponding degree values ​​of the processing difficulty and the idle time of the machine tool are calibrated as jh, js, sc, dz and xz respectively, and after normalization, they are entered into the following formula: To obtain the emergency value JXZ, and compare the obtained product emergency value with the tools that all auditors are reviewing and the corresponding product preset emergency values, obtain the preset emergency value lower than the product emergency value JXZ and the corresponding auditor and mark them, calculate the difference between the product emergency value JXZ and the marked preset emergency value, compare the obtained several differences to get the largest difference and the corresponding auditor, then send the audit data to the auditor's terminal, prompt the auditor to give priority to reviewing the tool corresponding to the product, and obtain the auditor's inspection receipt to determine whether the tool identification is correct.