Multi-process autonomous control method and system for factory of machining calipers based on internet of things

Abstract

The application discloses a kind of multi-process autonomous control method and system of factory based on Internet of Things, and the application relates to the technical field of autonomous control method, determine multiple surface treatment projects based on the identification of multistage surface treatment system, determine the multi-process autonomous control event of caliper in surface treatment stage according to the project content of multiple surface treatment projects, corresponding surface treatment part and the overall form of caliper, improve the accuracy of multi-process autonomous control event.Therefore, determine polishing event according to abnormal polishing area and subsequent polishing process;Determine multiple surface abnormal features according to polishing event and corresponding overall form, determine the surface autonomous processing event of caliper according to multiple surface abnormal features, surface processing requirements of caliper and multiple fine processing tools;Based on primary polishing event, polishing event and surface autonomous processing event, construct the multi-process autonomous control system of caliper.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of autonomous control method, and particularly relates to a multi-process autonomous control method and system for a factory processing calipers based on Internet of Things. BACKGROUND

[0002] With the development of science and technology, factories gradually move towards the stage of automation, and the workpieces to be processed are transported in the production line of the factory and sequentially pass through each process of the production line. In the prior art, the calipers sequentially pass through multiple processes of the production line, collect multiple processing data of the calipers in different processes, determine the corresponding processing surface according to the multiple processing data, and determine the corresponding surface quality based on the detection of the processing surface. However, the autonomous control of the primary polishing event, the polishing event and the surface autonomous processing event is ignored, and the multi-process autonomous control cannot be realized, which affects the accuracy of the multi-process autonomous control system. SUMMARY

[0003] The present application provides a multi-process autonomous control method and system for a factory processing calipers based on Internet of Things.

[0004] The present application provides a multi-process autonomous control method for a factory processing calipers based on Internet of Things, which comprises: processing the calipers in the factory in multiple processes, collecting the surface body processing image of the calipers based on Internet of Things and camera, determining each complete curved surface feature according to the recognition of the surface body processing image; determining the corresponding curved surface processing route according to the feature position of each complete curved surface feature and the overall shape of the calipers; determining multiple key processing nodes according to the curved surface processing route and the current posture of the calipers; determining a multi-level surface treatment system according to the multiple key processing nodes and the process table of the calipers, determining multiple surface treatment items based on the recognition of the multi-level surface treatment system, determining the multi-process autonomous control event of the calipers in the surface treatment stage according to the item content of the multiple surface treatment items, the corresponding surface treatment part and the overall shape of the calipers; determining the primary polishing event based on the multi-process autonomous control event; determining the abnormal polishing area of the calipers according to the recognition of the primary polishing event, determining the polishing event according to the abnormal polishing area and the subsequent polishing process; determining multiple surface abnormal features according to the polishing event and the corresponding overall shape, determining the surface autonomous processing event of the calipers according to the multiple surface abnormal features, the surface processing requirements of the calipers and multiple fine processing tools; constructing the multi-process autonomous control system of the calipers based on the primary polishing event, the polishing event and the surface autonomous processing event.

[0005] The embodiment of the present application provides a multi-process autonomous control system of a factory based on Internet of Things, which is applied to the multi-process autonomous control method of the factory based on Internet of Things for processing calipers, and comprises:

[0006] A complete curved surface feature module is used for multi-process processing of the caliper in the factory, and a face body processing image of the caliper is collected based on Internet of Things and a camera, and each complete curved surface feature is determined according to recognition of the face body processing image.

[0007] A key processing node module is used for determining a corresponding curved surface processing route according to a feature position of each complete curved surface feature and an overall form of the caliper, and determining a plurality of key processing nodes according to the curved surface processing route and a current posture of the caliper.

[0008] A multi-process autonomous control event module is used for determining a multi-level surface treatment system according to the plurality of key processing nodes and a process table of the caliper, determining a plurality of surface treatment items based on recognition of the multi-level surface treatment system, and determining a multi-process autonomous control event of the caliper in a surface treatment stage according to item content of the plurality of surface treatment items, a corresponding surface treatment part and the overall form of the caliper.

[0009] A polishing event module is used for determining a primary polishing event based on the multi-process autonomous control event, determining an abnormal polishing area of the caliper according to recognition of the primary polishing event, and determining a polishing event according to the abnormal polishing area and a subsequent polishing process.

[0010] A multi-process autonomous control system module is used for determining a plurality of surface abnormal features according to the polishing event and the corresponding overall form, determining a surface autonomous processing event of the caliper according to the plurality of surface abnormal features, surface processing requirements of the caliper and a plurality of fine processing tools, and constructing a multi-process autonomous control system of the caliper based on the primary polishing event, the polishing event and the surface autonomous processing event.

[0011] Compared with the prior art, the present application has the following advantages:

[0012] In the embodiment of the present application, the method in the embodiment of the present application is used to determine a multi-level surface treatment system according to a plurality of key processing nodes and a process table of a caliper, determine a plurality of surface treatment items based on recognition of the multi-level surface treatment system, determine a multi-process autonomous control event of the caliper in a surface treatment stage according to item content of the plurality of surface treatment items, a corresponding surface treatment part and an overall form of the caliper, introduce the multi-level surface treatment system, and compatibly consider the overall form of the overall form of the caliper, the item content of the plurality of surface treatment items and the corresponding surface treatment part, thereby improving the accuracy of the multi-process autonomous control event.

[0013] Therefore, the primary polishing event is determined based on the multi-process autonomous control event, the abnormal polishing area of the caliper is determined according to the identification of the primary polishing event, the polishing event is determined according to the abnormal polishing area and the subsequent polishing process, the multiple surface abnormal features are determined according to the polishing event and the corresponding overall morphology, the surface autonomous processing event of the caliper is determined according to the multiple surface abnormal features, the surface processing requirements of the caliper and the multiple fine processing tools, the multi-process autonomous control system of the caliper is constructed based on the primary polishing event, the polishing event and the surface autonomous processing event, the polishing event is introduced, and the primary polishing event, the polishing event and the surface autonomous processing event are further autonomously controlled, the overall consideration of the primary polishing event, the polishing event and the surface autonomous processing event is realized, and the precision of the multi-process autonomous control system is improved. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 is a flowchart of the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0015] Figure 2 is a flowchart of step S11 in the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0016] Figure 3 is a flowchart of step S12 in the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0017] Figure 4 is a flowchart of step S13 in the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0018] Figure 5 is a flowchart of step S14 in the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0019] Figure 6 is a flowchart of step S15 in the multi-process autonomous control method of the factory of the processing caliper based on the Internet of Things in the embodiment of the application.

[0020] Figure 7 is a structural composition schematic diagram of the multi-process autonomous control system of the factory based on the Internet of Things in the embodiment of the application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application.

[0022] Please refer to Figures 1 to 7A multi-process autonomous control method of a factory processing calipers based on the Internet of Things, applied to a multi-process autonomous control scene; the multi-process autonomous control method of the factory processing calipers based on the Internet of Things comprises:

[0023] Step S11: The calipers are processed in multiple processes in the factory, the surface body processing images of the calipers are collected based on the Internet of Things and the camera, and each complete curved surface feature is determined according to the recognition of the surface body processing images;

[0024] Step S12: The corresponding curved surface processing route is determined according to the feature position of each complete curved surface feature and the overall shape of the calipers; and a plurality of key processing nodes are determined according to the curved surface processing route and the current posture of the calipers;

[0025] Step S13: A multi-level surface treatment system is determined according to the plurality of key processing nodes and the process table of the calipers, a plurality of surface treatment items are determined based on the recognition of the multi-level surface treatment system, and a plurality of multi-process autonomous control events of the calipers in the surface treatment stage are determined according to the item content of the plurality of surface treatment items, the corresponding surface treatment part and the overall shape of the calipers;

[0026] Step S14: A primary polishing event is determined based on the multi-process autonomous control event; an abnormal polishing area of the calipers is determined according to the recognition of the primary polishing event, and a polishing event is determined according to the abnormal polishing area and the subsequent polishing process;

[0027] Step S15: A plurality of surface abnormal features are determined according to the polishing event and the corresponding overall shape, a surface autonomous processing event of the calipers is determined according to the plurality of surface abnormal features, the surface processing requirements of the calipers and a plurality of fine processing tools; and a multi-process autonomous control system of the calipers is constructed based on the primary polishing event, the polishing event and the surface autonomous processing event.

[0028] Reference Figure 2 In step S11, the specific steps are:

[0029] S111: The calipers are sequentially subjected to multiple processes in the factory under the driving of the conveying line and are processed in multiple processes; a plurality of cameras are distributed in corresponding processes and communicate with the Internet of Things, and the surface body processing process of the calipers is monitored in real time, and the surface body processing images of the calipers are collected based on the Internet of Things and the camera;

[0030] S112: A plurality of surface body processing features and remaining processing areas are determined according to the recognition of the surface body processing images, and corresponding remaining processing parts are determined based on the detection of the remaining processing areas; the plurality of surface body processing features gradually extend with the processing of the remaining processing parts to form each complete curved surface feature.

[0031] In the embodiments of the present application, each passing caliper is given a unique identity by the equipped RFID identification device, realizing accurate tracing throughout the whole process; the conveying line does not run at a constant speed, and its intelligence lies in its ability to automatically adjust the conveying speed according to the actual processing state of each downstream process (such as equipment load, task completion degree), and this dynamic response mechanism ensures flexible connection and efficient cooperation between processes.

[0032] The system deploys a multi-angle shooting array composed of 2 to 4 high-definition industrial cameras at each key processing procedure, ensuring that the processing details of the caliper are captured without dead angles; these cameras are not fixedly installed, but are installed on adjustable brackets, which can automatically and quickly adjust to the optimal shooting angle and distance according to the caliper model information recognized by the system, ensuring that different models of products can all obtain high-quality images.

[0033] The system adopts an advanced edge computing and Internet of Things combination mode; all cameras communicate with the central Internet of Things platform through high-speed industrial Ethernet or 5G network; but unlike the traditional method of uploading all raw data to the cloud, the system performs preliminary image processing and analysis at the camera end (edge side), such as image compression, feature extraction, etc., greatly reducing the amount of data that needs to be transmitted, reducing network bandwidth pressure, and improving response speed; the Internet of Things platform acts as the "brain", receiving real-time data from each process camera after preliminary processing, performing more in-depth centralized analysis, model training and long-term storage, realizing a closed loop from data acquisition to intelligent decision-making.

[0034] Uninterrupted real-time monitoring of the caliper surface processing process; it continuously collects images at a high frequency of 10 to 30 frames per second, ensuring that every subtle dynamic of the processing process can be continuously recorded; in order to intelligently find abnormalities from the massive images, the system uses an image difference algorithm, which can detect any subtle changes in the processing area in real time and accurately, such as tool scratches, material excess or loss, etc.; further, the system builds a processing process digital twin model that is completely synchronized with the physical production line based on the collected real-time data, and the operator can intuitively see the real-time processing state of each caliper in the virtual environment, realizing deep integration and synchronous monitoring of virtual and reality.

[0035] The image data collected by the system is extremely rich and structured; it not only includes raw images without processing, but also high-quality images pre-processed by algorithms such as noise reduction and enhancement, providing double protection for subsequent accurate analysis; in addition, the system also introduces multispectral imaging technology, which can capture spectral information of different wavebands and obtain deep information such as material composition, coating thickness and internal texture on the surface of the caliper that cannot be distinguished by the naked eye; "surface-body processing image" refers to the image data collected by the high-definition industrial camera installed in each process of the caliper multi-process machining, which reflects the processing state and details of the caliper surface. These images not only include the original pictures, but also the pre-processed and enhanced images, which can clearly show the processing characteristics, texture, dimensional accuracy and defects of the caliper surface; at the same time, all these image data will not exist in isolation, but will be packaged and stored with corresponding process information, time stamp accurate to milliseconds, running parameters of the current processing equipment and other key metadata, forming a complete, traceable and analyzable data archive.

[0036] Further, the system performs in-depth analysis on the collected caliper surface-body processing images to determine multiple key processing features. This process mainly relies on advanced deep learning algorithms such as convolutional neural networks or YOLO models, which can automatically extract core features such as the outline, edge and hole of the surface-body from complex image backgrounds; to ensure the accuracy and standardization of identification, the system pre-establishes a detailed caliper curved surface feature library, which not only contains standard curved surface three-dimensional models, but also labels all key feature points and feature lines.

[0037] At the same time, the system will first use image segmentation technology to intelligently divide the entire surface of the caliper into multiple feature areas with specific geometric meanings, such as planes, transition curves, holes and grooves; then, the system will accurately calculate the geometric parameters of each feature area, such as curvature, normal vector, area and perimeter, and compare these real-time calculated features with the standard models in the feature library with high precision, to instantly determine the current processing state.

[0038] The system focuses on intelligent recognition and planning of the "remaining machining part". It performs pixel-level fine comparison between the real-time collected image of the caliper that has not yet been completed and the pre-set standard completed model through image difference algorithm, thereby accurately identifying all areas that have not yet met the machining requirements and explicitly marking them as "remaining machining areas". In order to achieve fine machining, the system intelligently classifies these remaining areas, such as dividing them into rough machining allowance areas, fine machining allowance areas, or surface treatment areas, etc. Furthermore, the system quantitatively calculates key parameters such as the machining allowance (i.e. the thickness of material to be removed) and the estimated surface roughness of each remaining machining area, and based on the size and precision requirements of these parameters, automatically matches and determines the optimal machining process route and required tools from the process database, providing precise instructions for the next stage of machining.

[0039] The entire machining process is not fixed, but a dynamic evolution process. The system establishes a dynamic feature tracking mechanism to monitor the dynamic changes in the boundary between the machined features and the unprocessed areas as material is removed. It uses optical flow or feature point matching algorithms to continuously track the morphological evolution of the feature areas during continuous machining, just like a precise observer. With each cutting of the tool, the system updates the feature boundary in real time, achieving "dynamic extension" of the machining features. To be forward-looking, the system also builds a feature growth model that can predict the state of the features after the next machining step based on the current machining state and the set process parameters. Through a precise feedback control mechanism, the system can dynamically adjust the subsequent machining parameters (such as feed speed, spindle speed, etc.) based on the deviation between the prediction and the actual state, to ensure that all features are continuously, consistently, and accurately formed, and are spliced into complete and designed curved surface features.

[0040] Reference Figure 3 In step S12, the specific steps are:

[0041] S121: Collect each complete curved surface feature, determine the feature position of each complete curved surface feature based on the detection of each complete curved surface feature, at the same time, collect the overall shape of the caliper and the corresponding surface machining requirements, determine the curved surface machining route of the caliper according to the feature position of the plurality of complete curved surface features, the overall shape of the caliper and the corresponding surface machining requirements;

[0042] S122: Real-time monitoring of the caliper and collecting a plurality of posture parameters of the caliper, determining the current posture of the caliper according to the plurality of posture parameters of the caliper, and determining the corresponding curved surface machining tool according to the current posture of the caliper, the corresponding curved surface machining content and the machining tool database; determining a plurality of key machining nodes of the caliper based on the curved surface machining route of the caliper, the current posture of the caliper and the corresponding curved surface machining tool.

[0043] In the embodiments of the present application, a high-precision three-dimensional scanner or a multi-view vision system is used to scan the caliper comprehensively and without dead angles, collecting a large amount of point cloud data on the surface. These data points from different angles, like scattered puzzle pieces, need to be accurately aligned and fused through a point cloud registration algorithm (such as the ICP iterative closest point algorithm) to piece together a complete and unified point cloud model.

[0044] Using professional surface reconstruction algorithms (such as Poisson reconstruction or Delaunay triangulation), these discrete point cloud data are converted into continuous and smooth surface mathematical representations; to eliminate noise and minor flaws generated during the scanning process, the system also performs smoothing and filtering on the reconstructed surface, ensuring the purity and high quality of the data; it is also a crucial step to accurately register the surface model reconstructed by scanning with the original CAD design model, establishing a unified coordinate system correspondence, ensuring that all subsequent measurements, analyses, and planning are based on the same spatial reference.

[0045] Specifically, in the actual operation of the factory, for the A type automobile brake caliper, the technical personnel used a high-precision blue light three-dimensional scanner; the scanner scanned the caliper from 8 preset different angles in a circumferential manner, and each angle could collect about 2 million accurate three-dimensional coordinate points, a total of 16 million data points; the system immediately starts the ICP algorithm, like a puzzle master, accurately splices and aligns the 8 groups of point cloud data to form a complete and gapless A type caliper point cloud model.

[0046] The system uses Poisson reconstruction algorithm to convert this point cloud model into a continuous and editable surface representation, and applies Laplace smoothing technique to effectively remove minor surface noise generated during the scanning process; the system automatically registers the reconstructed three-dimensional model with the CAD design model of the A type caliper, establishing a globally unified coordinate system.

[0047] The system will take the geometric center of the caliper or a designated reference point as the origin to establish a global coordinate system; through feature extraction algorithm, it automatically identifies and calculates the accurate position and pose of each independent surface feature (such as plane, hole, surface, etc.) in the coordinate system, which includes calculating the geometric center, boundary contour, and a series of key feature points (such as center, corner, etc.) of each feature; more importantly, the system analyzes and determines the relative position relationship and spatial constraints between these features, such as distance, angle, parallelism, coaxiality, etc.; all these accurate position and pose information, including coordinates, normal vectors, curvatures, dimensions, etc., will be integrated into a feature position matrix and stored in a special feature position database.

[0048] Specifically, for the A-type caliper, the system establishes a global coordinate system with its geometric center as the origin; through the feature extraction algorithm, the system accurately identifies and calculates the position parameters of its five main curved surface features: the piston hole surface, with its center at the origin (0, 0, 0), the normal vector vertically upward (0, 0, 1), the diameter 60mm, the depth 40mm; the brake pad contact surface, with its center coordinates (30, 0, 0), the normal vector horizontally to the right (1, 0, 0), the size of the rectangular 80mm x 60mm; the mounting flange surface, with its center coordinates (-25, 0, 0), the normal vector horizontally to the left (-1, 0, 0), the ring shape, the outer diameter 100mm, the inner diameter 80mm. The left connecting curved surface, with its center coordinates (0, 35, 0), the normal vector forward (0, 1, 0), the maximum curvature 0.05mm⁻¹; the right connecting curved surface, with its center coordinates (0, -35, 0), the normal vector backward (0, -1, 0), the maximum curvature also 0.05mm⁻¹.

[0049] The system further calculates the key relative positions between them, such as the distance from the piston hole surface to the brake pad contact surface is 30mm, to the mounting flange surface is 25mm, and the distance from the connecting curved surface on both sides to the piston hole surface is 35mm. These spatial relationship data accurate to the micron level are all stored in the feature position database, constituting the geometric constraint blueprint for machining route planning.

[0050] The system will again use the three-dimensional scanning data to extract the overall macroscopic morphological parameters of the caliper, such as total length, total width, total height, weight, etc.; the system will automatically retrieve its corresponding surface machining requirements from the process database according to the type of the caliper, which includes the precision level of different feature surfaces (such as IT6, IT7), the surface roughness (such as Ra0.8, Ra1.6), and various geometric tolerances (such as flatness, cylindricity, parallelism, etc.); In addition, the system will also collect or retrieve the material property information of the caliper, such as material grade, heat treatment state, hardness, toughness, etc., which directly affect the selection of cutting tools and cutting parameters; the system will obtain the functional requirements of the caliper, such as a certain surface needs to realize sealing fit, a certain surface needs to be uniformly stressed, etc.; all these information are integrated and correlated by the system, forming a complete machining requirement document of the caliper with detailed content and clear structure.

[0051] Specifically, the system extracts the overall shape data of the A-type caliper from the three-dimensional model: total length 120 mm, width 90 mm, height 70 mm, and weight about 1.2 kg; then, the detailed machining requirements of this type are retrieved from the process database: the material is aluminum alloy ADC12, which is heat treated to T6, and the hardness is HB90; the specific requirements of each feature surface are: the piston hole surface needs to reach IT6 level precision, the surface roughness Ra0.8, and the cylindricity error is not more than 0.01 mm; the brake pad contact surface and the mounting flange surface need to reach IT7 level precision, the surface roughness Ra1.6, and the flatness and parallelism requirements are within 0.02 mm and 0.015 mm respectively; the two side connecting curved surfaces are IT8 level precision, the surface roughness Ra3.2, and the profile tolerance is 0.03 mm; at the same time, the system also clearly defines its functional requirements: the piston hole surface must form a reliable sealing fit with the piston, the brake pad contact surface needs to ensure uniform contact with the brake pad to transmit braking force, and the mounting flange surface needs to be firmly connected with the frame; all these information is integrated into a complete machining requirement document, providing comprehensive constraints and targets for subsequent process planning.

[0052] Based on all the information obtained in the previous stage (geometric features, position relationships, machining requirements), an optimal machining route is planned; the system usually uses graph theory or intelligent optimization algorithms (such as genetic algorithm) to solve it; when planning, the system will consider multiple key factors, such as machining accuracy (reference selection and transmission), machining efficiency (minimum number of processes, shortest time), tool life (cutting parameter optimization), etc.

[0053] The specific content of planning includes: determining the machining sequence of each feature (rough first, then fine, face first, then hole, reference first, then others), selecting the most suitable machining method for each feature (such as milling, turning, grinding, polishing), planning the connection and positioning method between processes (how to use the already processed surface as the reference for subsequent processing), and setting the optimal process parameters (cutting speed, feed rate, cutting depth) for each step; the system will generate an intuitive and complete machining route diagram, which clearly shows the entire machining process from blank to finished product in a graphical and data-based manner.

[0054] Specifically, the system uses genetic optimization algorithm based on its geometric features, position relationships and machining requirements, and after tens of thousands of iterations, it plans an optimal machining route; the machining sequence is determined as:

[0055] 1) Machining the mounting flange surface - it is chosen as the first process because it is the largest plane, most suitable as a stable positioning reference for all subsequent processing; turning is used, and the process parameters are set to cutting speed 200 m / min, feed rate 0.1 mm / r, and cutting depth 1.5 mm;

[0056] 2) Machining the piston hole face - using the already machined mounting flange face as a positioning reference to ensure the hole's positional accuracy; using boring processing, cutting speed 150 m / min, feed rate 0.08 mm / r, cutting depth completed in three steps (1.5 mm, 1.0 mm, 0.5 mm) to ensure the hole's accuracy and surface quality;

[0057] 3) Machining the brake pad contact face - using the two already machined high-precision faces (the mounting flange face and the piston hole face) as positioning references (two faces and one pin) to ensure their positional and directional accuracy; using milling processing, cutting speed 180 m / min, feed rate 0.1 mm / tooth, cutting depth 1.2 mm; 4) Machining the two connecting curved surfaces - using the three-face positioning system formed by the three already machined faces, a five-axis machining center is used to ensure the continuity and contour accuracy of the complex curved surfaces; cutting speed 160 m / min, feed rate 0.08 mm / tooth, cutting depth 1.0 mm.

[0058] The connection method between processes is also carefully designed: from the mounting flange face to the piston hole face, one-face-one-pin positioning is used; from the piston hole face to the brake pad contact face, it is upgraded to two-face-one-pin positioning; when machining the connecting curved surfaces, the most stable three-face positioning is used; a detailed machining route map is generated, including all machining sequences, methods, parameters, positioning methods, and key quality control points, providing a scientific and complete guidance scheme for the efficient and high-precision production of the A-type caliper, perfectly balancing product quality and production efficiency.

[0059] Further, the system deploys various high-precision sensors at key parts of the caliper's fixture and machining equipment, including inclination sensors, gyroscopes, and accelerometers, to measure subtle angular changes and vibrations; simultaneously, a visual measurement system tracks pre-set marker points or matches inherent features on the caliper to achieve non-contact position tracking; to further enhance precision, laser trackers or high-precision optical measurement systems are introduced to obtain the caliper's complete six-degree-of-freedom (three translations and three rotations) attitude information in space; all these sensors are connected into a data acquisition network through high-speed industrial Ethernet, continuously collecting raw attitude data at a frequency of up to 100 Hz; before entering the analysis stage, these raw data undergo a series of complex filtering and data fusion algorithms to eliminate noise and compensate for errors, thereby significantly improving the accuracy and reliability of the measurement results.

[0060] Specifically, three inclination sensors with a precision of 0.001° and two six-axis accelerometers are installed on the caliper's special fixture to monitor its inclination state and slight vibrations in real time; around the machining area, four 5 million-pixel industrial cameras form a visual measurement system, which accurately tracks the position changes of the caliper's surface by identifying four 5mm-diameter circular marker points pre-set on the surface.

[0061] In addition, the five-axis machining center itself integrates a laser tracker with a measurement accuracy of ±0.005 mm, which can provide sub-millimeter level position feedback. These sensors form a high-speed data acquisition network through industrial Ethernet, synchronously collecting data at a frequency of 100 Hz, including the displacement of the caliper in X, Y, Z three axes (accuracy ±0.01 mm), the rotation angle around the three axes (accuracy ±0.005°), and the vibration acceleration of the three axes (accuracy ±0.001g); The massive raw data collected will immediately be processed by wavelet denoising and Kalman filtering algorithms, providing clean and reliable data sources for subsequent attitude calculation.

[0062] The system usually uses advanced data fusion algorithms such as Kalman filtering or particle filtering, which can effectively handle the uncertainty of sensor data and output an optimal attitude estimate; The calculation result not only includes the real-time position coordinates and rotation angles of the caliper, but also systematically evaluates the uncertainty and potential error range of this measurement, providing risk basis for subsequent decision-making; In addition, the system also establishes an attitude prediction model based on the current motion state of the caliper (such as speed, acceleration), to predict its attitude in a very short time, which is crucial for forward-looking machining control; All these information is integrated into a detailed real-time attitude report, which clearly shows the key dynamic parameters such as position, angle, speed and acceleration of the caliper.

[0063] Specifically, at a certain processing time, the system uses the extended Kalman filter algorithm to fuse all data from the tilt sensor, accelerometer, vision system and laser tracker to accurately calculate the current attitude of the A-type caliper; The report shows that its position is X=125.32 mm, Y=78.45 mm, Z=62.18 mm; The attitude is rotated around the X-axis by 0.125°, around the Y-axis by -0.083°, and around the Z-axis by 0.056°; At the same time, its motion speed is 0.12 mm / s in the X direction, -0.08 mm / s in the Y direction, and 0.05 mm / s in the Z direction; The vibration acceleration remains at a very low level.

[0064] The system also gives the confidence interval of the measurement: the position error is within ±0.015 mm, and the angle error is within ±0.008°; Based on the current motion trend, the attitude prediction model further predicts that after 0.5 seconds, the position of the caliper will change to X=125.38 mm, Y=78.41 mm, Z=62.20 mm, and the attitude change will be less than 0.01°, this real-time attitude report is intuitively displayed on the central control interface.

[0065] On the basis of accurately grasping the current posture of the caliper, the system needs to intelligently select the most suitable tool for the next machining operation, and this process relies on a large machining tool database that stores detailed information on hundreds of tools, including geometric parameters (diameter, blade length, helix angle, etc.), material properties (cemented carbide, high-speed steel, coating, etc.), and optimal application range (machining material, process type, precision level, etc.).

[0066] The system will preliminarily screen a series of candidate tools from the database according to the current machining surface characteristics (such as curvature, roughness requirements) and the material of the caliper; the system will perform a critical accessibility and interference analysis: it will simulate the real-time posture of the caliper and the geometric model of the candidate tools in a virtual environment, simulate the movement of the tool along the predetermined path, and check whether it will collide with the caliper or fixture; under the premise of meeting all geometric constraints, the system will evaluate the machining efficiency and expected life of each candidate tool, select the one with the best overall performance, and determine the precise installation parameters (such as extension length) and optimized cutting parameters (such as spindle speed, feed speed) of the tool.

[0067] Specifically, assuming that a model A caliper is currently machining its left connecting surface, the system starts the tool selection process based on the real-time posture report and machining requirements; the characteristics of this surface are a maximum curvature of 0.05 mm⁻¹, a surface roughness requirement of Ra3.2, and the material is aluminum alloy ADC12; the system selects tools suitable for aluminum alloy finishing from the tool database; then, considering the current slight tilt of the caliper by 0.125° on the Y-axis, the system performs a detailed accessibility analysis, simulates the machining path of multiple tools in the current posture, and excludes options that interfere; after comprehensive evaluation, the system selects a TiAlN-coated cemented carbide ball end mill with a diameter of 8 mm; and determines its optimal working parameters: tool extension length 35 mm, spindle speed 13000 r / min (slightly higher than the standard value to compensate for the change in cutting force due to the tilt), feed speed 1800 mm / min, and cutting depth 1.0 mm, which ensures both machining accuracy and efficiency.

[0068] Based on the pre-planned machining route and the real-time posture of the caliper, identify the "key machining nodes" on the machining path that have a decisive influence on the final quality, which usually include: tool initial cutting point, point where curvature changes significantly, point where cutting force changes abruptly, and finishing area that requires the highest precision control, etc.; the system will clearly define the precise spatial position of each key node, the posture requirements that the tool needs to reach, and the special control strategies that need to be executed at that point (such as reducing the feed speed, enabling power monitoring, etc.); at the same time, the system will also clearly define the items that need to be detected at each node, such as cutting force, vibration, surface profile, etc., and generate a list of key machining nodes.

[0069] Specifically, for the machining of the left connecting curved surface of the A-type caliper, the system plans 5 key machining nodes based on its machining route and current pose. For example, node 1 (initial contact point), the system requires the tool to cut in at a specific angle and reduce the feed speed to 800 mm / min, while monitoring the cutting force to ensure smooth start. Node 3 (cutting force mutation point), the system predicts that the cutting load will increase significantly at this point, so it requires the tool pose to be adjusted to an angle less than 3° with the normal of the curved surface, and the feed speed to be reduced to 1200 mm / min, while the spindle power and temperature monitoring are turned on. In actual machining, when the system detects a 20% increase in cutting force at node 3, it immediately executes the preset speed reduction and pose adjustment strategy, successfully avoiding potential surface defects. At node 4 (precision control point), the system activates the finishing mode, reduces the feed speed to 1000 mm / min, and integrates online 3D scanning function. When the scan finds that the local profile deviation reaches 0.035 mm (exceeding the tolerance requirement of 0.03 mm), the system immediately plans a compensation path for secondary machining, finally correcting the profile of the area to 0.028 mm, perfectly meeting the design requirements.

[0070] Reference Figure 4 In step S13, the specific steps are:

[0071] S131: Determine the node positions and corresponding node morphologies of the plurality of key machining nodes based on the detection of the plurality of key machining nodes, and determine the process table of the caliper based on the type of the caliper, the corresponding overall morphology and the factory machining database, and determine the multi-level surface treatment system according to the node positions, corresponding node morphologies of the plurality of key machining nodes and the process table of the caliper;

[0072] S132: Determine the plurality of surface treatment items according to the identification of the multi-level surface treatment system, and mark the item contents of the plurality of surface treatment items, and determine the corresponding surface treatment part according to the item contents of the plurality of surface treatment items and the current pose of the caliper;

[0073] S133: Determine the first autonomous control coefficient according to the item contents of the plurality of surface treatment items and the overall morphology of the caliper, determine the second autonomous control coefficient according to the surface treatment part corresponding to the plurality of surface treatment items and the overall morphology of the caliper, and determine the multi-process autonomous control event of the caliper in the surface treatment stage based on the first autonomous control coefficient, the second autonomous control coefficient and the autonomous control event mapping relationship of the surface treatment stage.

[0074] In the embodiments of the present application, a high-precision measurement system such as a laser scanner or a three-coordinate measuring machine is used to conduct non-contact or contact measurement on each key machining node to determine its absolute coordinates in three-dimensional space with micron-level precision; at the same time, through a high-resolution industrial camera combined with an advanced image processing algorithm, the surface topography of the node area is analyzed in depth to obtain morphological features including surface roughness, tool mark texture, microscopic defects, etc.; the system will further calculate the geometric parameters of each node such as local curvature, normal vector direction, etc. to comprehensively describe its geometric properties; all these data - position coordinates, morphological features and geometric parameters - are systematically stored in a node morphology database and a intuitive node position-morphology mapping relationship diagram is generated to clearly show the distribution of each node in space and its corresponding morphological quality condition.

[0075] The system will retrieve the standard process flow corresponding to the A-type caliper from the central product database, which provides a reference framework; the system will fine-tune the standard process according to the overall morphological parameters of this batch of calipers (such as size, weight, structural complexity, etc.); more importantly, the system will access the factory's machining database in real time to obtain the current production resource status, including which equipment is available, tool inventory, skill level of operators, etc.; based on these multi-dimensional information, the system will use intelligent optimization algorithms to dynamically adjust the process sequence, machining content, equipment allocation to balance the machining time of each process, maximize equipment utilization, and ensure smooth production flow; the output process table is a detailed job instruction book.

[0076] The system will analyze the position distribution and morphological features of the key nodes to identify which are the key areas that bear core functions and require the highest quality, which are the less important areas, and which are the areas that only require basic processing; according to the overall surface treatment requirements set in the process table, the entire treatment process is divided into multiple levels, such as basic level, standard level and fine level; after determining the level, the system will combine the material properties of the caliper (such as the grade of aluminum alloy, heat treatment state) and functional requirements (such as wear resistance, corrosion resistance, sealing performance) to accurately set technical parameters for each level of treatment, including treatment method, tool selection, treatment time, quality target, etc.; these different levels of treatment methods are integrated into a hierarchical and logically clear multi-level treatment system, which usually includes three levels of pretreatment (to solve basic problems), main treatment (to achieve core function improvement) and post-treatment (to provide protection and performance enhancement), and generates the corresponding system document as the basis for on-site operation.

[0077] Specifically, the technician used a high-precision laser scanner to conduct a comprehensive inspection of five key processing nodes on the caliper; the scanner obtained the three-dimensional coordinates of each node with an accuracy of 0.001 mm, while a high-resolution camera with 20 million pixels captured the surface micrograph; the inspection results showed that the surface quality of node 3 (position: 132.18 mm, 85.15 mm, 66.85 mm) was the worst, with a surface roughness of Ra2.0, and microscopic morphology analysis showed the presence of uneven cutting marks and slight vibration marks; in contrast, the quality of node 5 (position: 138.25 mm, 90.25 mm, 70.85 mm) was the best, with a surface roughness of only Ra1.2, and the morphology was uniform; the system stored these precise data and generated a mapping chart, which directly revealed the surface quality differences in different areas of the caliper.

[0078] For the A-type caliper, the system retrieved its standard process flow containing 20 procedures; then, combined with its overall form (total length 120 mm, weight 1.2 kg, medium structural complexity), it made preliminary adjustments; most importantly, the system evaluated the real-time status of the factory: three CNC machining centers were in good condition, but the polishing equipment load had reached 80%, the anodizing line was running normally, and there were five skilled workers and three apprentices on duty.

[0079] Based on this information, the system optimized the scheduling of the procedures; for example, considering the high load of the polishing equipment, the system moved some non-critical path procedures forward to leave a wider time window for the polishing procedure, and the generated procedure table (only the surface treatment-related part is listed) clearly specified: procedure 15 (deburring) needs to be completed within 5 minutes, with a quality standard of no visible burrs; procedure 16 (rough polishing) aims to reduce the surface roughness to below Ra1.6; and until procedure 20 (sealing treatment), to ensure that the final product meets the corrosion resistance standard.

[0080] For the A-type caliper, the system analysis found that nodes 1-3, located on the piston bore face and brake pad contact face, are the key working surfaces directly related to brake performance and sealing performance, and the initial quality is poor (especially node 3), which requires the highest level of processing; while nodes 4-5 are located on the mounting flange face, with good initial quality and relatively low functional requirements; based on this, the system establishes a three-level processing system: the first-level processing (basic level) covers 100% of the outer surface, mainly using mechanical deburring and rough grinding, the goal is to eliminate burrs and control the overall roughness to below Ra1.6; the second-level processing (standard level) is specifically for nodes 1-3 area, using fine grinding and preliminary polishing, the goal is to improve the roughness of these key surfaces to Ra1.0; the third-level processing (fine level) is for nodes 4-5 area, using fine polishing and anodizing, not only to achieve an extremely high smoothness of Ra0.8, but also to form a 15-20μm thick hardening oxide film to improve its wear resistance and corrosion resistance; the entire system from pretreatment (first level) to main processing (second and third levels) to closed processing (post-processing).

[0081] Further, a plurality of surface treatment items are determined according to the identification of the multi-level surface treatment system, and the item content of the plurality of surface treatment items is marked, and a corresponding surface treatment part is determined according to the item content of the plurality of surface treatment items and the current posture of the caliper.

[0082] At this time, the system will traverse the entire multi-level processing system, and according to the differences in processing level, target area and core method, it will be decomposed into multiple independent items; for example, a “second-level standard processing” will be decomposed into “fine grinding” and “polishing” two consecutive items; for each item, the system will explicitly assign its processing level (such as second level), processing area (such as nodes 1-3) and processing method (such as precision grinder).

[0083] At the same time, the system will define the logical relationship between these items, such as which must be executed in sequence, which can be parallel, thereby forming an optimized project network diagram; the system will generate a structured processing item list, including item number, name, level, area and other key attributes.

[0084] The system will define the specific content of each project in detail, such as "manually remove burrs and sharp edges on all external surfaces and edges of the caliper"; more importantly, the system will accurately mark various process parameters, including processing time (such as 5 minutes), operating pressure, equipment speed, working temperature, etc., to ensure the standardization and repeatability of the processing process; at the same time, the system will specify the specific tools, equipment and consumables required to perform the project, such as "stainless steel scraper, file, sandpaper (240 mesh)" or "belt sander (model DB-120)"; for quality control, the system will clearly define the quality standards and corresponding detection methods for each project, such as "no visible burrs, smooth edge transition" and "visual inspection + tactile inspection + magnifying glass (10 times) inspection"; all these detailed information is structured stored in the project content database, supporting real-time query, update and version management, forming a dynamic, digital process knowledge base.

[0085] The system will analyze the current pose of the caliper on the fixture (such as X-axis deflection of 0.5° and Y-axis deflection of 0.3° for A-type caliper), and combine it with its three-dimensional model to evaluate the accessibility and processing difficulty of each surface area; through motion simulation, the system can calculate a difficulty coefficient for each potential processing area (such as 0.3 for easy processing and 0.9 for very difficult processing); the system will match the project content defined in the second step with these surface areas with difficulty coefficients to determine which physical parts each processing project will act on; during the matching process, the system will also consider the geometric constraints and motion space of the processing tool to optimize the division of the processing area, ensuring that the tool can reach the target position without interference; the system will establish a one-to-one mapping relationship between the processing project and the surface processing part, and generate an intuitive "surface processing area map", which is usually presented in the form of three-dimensional rendering, with different colors representing different processing projects, and the depth of color or specific legends representing the difficulty of processing.

[0086] Specifically, for the A-type caliper, the system successfully decomposed and defined 6 specific surface treatment items based on its three-level surface treatment system, covering the whole process from basic preparation to final finishing; for example, the first-level treatment was decomposed into P1 (overall deburring treatment) and P2 (overall rough polishing treatment), both of which must be executed in sequence to remove burrs before overall polishing; the second-level treatment targets the key working surface (node 1-3 area), which is decomposed into P3 (fine polishing treatment) and P4 (polishing treatment), both of which must also be executed in sequence to ensure that the surface quality is gradually improved; the third-level treatment corresponds to P5 (node 4-5 area fine polishing treatment), which is an independent high-precision item; P6 (overall anodizing treatment) is arranged after all mechanical surface treatment is completed as a cross-level final treatment item; the system clearly defines the execution logic between them: P1→P2→(P3→P4)∥P5→P6, which means P5 can be performed in parallel with P3 and P4 to shorten the total working hours.

[0087] For the P3 item of the A-type caliper (node 1-3 area fine polishing treatment), the item content is marked in great detail; the treatment content is clearly defined as "precise polishing of the piston hole surface and brake pad contact surface"; the process parameters are strictly limited: grinding wheel size 240 mesh, speed 3000 r / min, feed speed 50 mm / min; the equipment used is specified as "precision polisher (model PM-240)", grinding wheel diameter 100 mm; the quality standard contains quantitative indicators: surface roughness ≤ Ra1.2, flatness ≤ 0.02 mm, no corrugation; the detection method is also specified as "surface roughness meter measurement + flatness measurement + microscope inspection"; similarly, for the P6 item (overall anodizing treatment), its chemical parameters (sulfuric acid concentration 180 g / L, temperature 20℃), electrical parameters (current density 1.5 A / dm²) and quality requirements (oxidation film thickness 15-20 μm, hardness ≥ HV300) are all accurately marked, and this level of detailed marking ensures that no matter who the operator is or on which equipment, consistent and standard-compliant results can be achieved.

[0088] The system analyzes the fixed posture and evaluates the processing difficulty of each area: the upper surface is the easiest (0.3), and the concave surface is the most difficult (0.9); then, the system accurately maps the six processing items to each part of the caliper; for example, the P3 item (node 1-3 area fine polishing processing) is further subdivided into three processing parts: the piston hole surface (difficulty 0.7), the brake pad contact surface center (difficulty 0.5), and the edge (difficulty 0.8); for the edge area with a difficulty of up to 0.8, the system optimizes the feed path of the polishing tool through simulation to ensure effective access; for the connecting surface in the P5 item (difficulty 0.9), the system specially specifies the use of a more flexible and smaller sponge wheel tool, and the generated surface processing area map clearly marks the high-difficulty connecting surface in the P5 item with deep red and the easily processed upper surface in the P2 item with light green, making the entire complex surface processing task clear at a glance.

[0089] Therefore, the first autonomous control coefficient is determined according to the project content of the plurality of surface processing projects and the overall shape of the caliper, the second autonomous control coefficient is determined according to the surface processing part corresponding to the plurality of surface processing projects and the overall shape of the caliper, and the multi-process autonomous control event of the caliper in the surface processing stage is determined based on the first autonomous control coefficient, the second autonomous control coefficient, and the autonomous control event mapping relationship of the surface processing stage. The overall consideration of the first autonomous control coefficient, the second autonomous control coefficient, and the autonomous control event mapping relationship of the surface processing stage ensures the accuracy of the multi-process autonomous control event of the caliper in the surface processing stage. At the same time, a multi-level surface processing system is introduced, which considers the overall consideration of the project content of the plurality of surface processing projects, the corresponding surface processing part, and the overall shape of the caliper, and improves the accuracy of the multi-process autonomous control event.

[0090] At this time, the system will deeply analyze each determined surface processing project, such as deburring, rough polishing, and fine polishing, and extract key process parameters (such as processing time, speed, and pressure) and strict technical requirements (such as surface roughness and flatness); the system will comprehensively evaluate the complexity of each project, which not only considers technical difficulty, but also considers process stability, parameter sensitivity, and operation precision requirements; for example, a simple deburring process has low complexity, while an anodizing process that requires accurate control of multiple parameter couplings has high complexity; the system will give each project a complexity score.

[0091] The system will consider the physical characteristics of the caliper itself, such as its overall geometry, material properties, and structural complexity (such as the number of concave surfaces, holes, etc.), as these factors directly affect the difficulty of process execution; the system will use a weighted scoring mathematical model to combine the complexity scores of all items with the influence coefficient of the overall shape of the caliper to obtain the first autonomous control coefficient, which is a value between 0 and 1, the higher the value, the more it needs highly intelligent autonomous control to ensure quality from the perspective of process content.

[0092] Specifically, when processing A-type calipers, the system analyzes 6 surface treatment items; P1 (deburring) has a simple technology, with a complexity score of only 0.3; while P6 (anodizing) involves multi-parameter precise control, with a complexity score as high as 1.0; at the same time, the system assesses the influence coefficient of A-type calipers (aluminum alloy material, medium complexity structure) on the processing process as 0.8; through weighted formula calculation, the system obtains the first autonomous control coefficient as 0.49, which indicates that from the perspective of process content, the surface treatment task of the caliper has a medium level of autonomous control demand, meaning that it cannot rely completely on fixed procedures and needs a certain degree of intelligent regulation to cope with process uncertainties.

[0093] The system will analyze the spatial distribution characteristics of each processing part (i.e. the specific surface area to be processed) in detail, including the size of the area, the complexity of the geometric shape (whether it is a plane or a complex curved surface), and most importantly, the accessibility of the tool; for example, a flat and open outer surface has good accessibility; while a deep internal hole or a narrow corner has poor accessibility, and the processing difficulty is naturally high.

[0094] The system will evaluate the geometric complexity and processing difficulty of each processing part; in addition, the system will also consider the current fixed posture of the caliper (such as whether there is a deflection, tilt), as the posture will directly affect the ease of tool access to the workpiece from different angles; the system will use a geometric complexity index mathematical model to combine the spatial feature scores of all processing parts with the influence coefficient of the current posture of the caliper to obtain the second autonomous control coefficient; similarly, this coefficient is also between 0 and 1, the higher the value, the greater the challenge faced in the spatial execution level, the more it needs intelligent path planning, posture adjustment, obstacle avoidance, and other autonomous control capabilities.

[0095] Specifically, the system evaluates the spatial challenges of each processing part; P1 (deburring) is processing the edge, but most of it is easy to access, and the geometric complexity score is 0.5; while the connecting surface in P5 (fine polishing) has extremely complex shape and needs multi-axis linkage to process, the geometric complexity score is as high as 0.95; at the same time, the influence coefficient of the current posture of the caliper (X-axis deflection 0.5°, Y-axis deflection 0.3°) on the processing process is 0.7; through the geometric complexity index method, the system obtains the second autonomous control coefficient as 0.50, which is slightly higher than the first coefficient, clearly indicating that in the surface processing of the A-type caliper, how to make the tool accurately, safely and efficiently reach those complex geometric surfaces is a more prominent challenge than controlling the process parameters itself, and the demand for autonomous control has reached a medium-high level.

[0096] The system will establish a mapping table, which is like a decision matrix, defining what level of control strategy should be corresponding to the different combination intervals of the first and second coefficients (for example, when both coefficients are low, only basic monitoring is needed; when both coefficients are high, the highest level of intelligent control is needed); the system will determine the level of overall control demand according to the calculated two coefficients; the system will intelligently select a series of appropriate control events from a pre-set control event library, these events are pre-programmed control modules that can solve specific problems, such as adaptive parameter adjustment (automatically correcting parameters when quality is not up to standard), tool path optimization (automatically planning new paths when there is a risk of collision), multi-process collaborative control (automatically compensating when upstream processes affect downstream) and so on.

[0097] The system will define the trigger conditions (such as quality deviation exceeding 10%), specific control actions (such as increasing speed) and expected effects (such as quality returning to normal) for each selected event; the system will generate an ordered autonomous control event sequence, clearly defining the logical relationship between these events (such as which has the highest priority, which can be executed in parallel), forming an intelligent control network that runs through the entire multi-process machining process.

[0098] Specifically, for the A-type caliper, the combination of the first coefficient (0.49) and the second coefficient (0.50) is determined as medium overall control demand through the intelligent control event mapping table, corresponding to the intelligent control event type; at the same time, the intelligent control event mapping table is shown in Table 1:

[0099] Table 1 Intelligent Control Event Mapping Table

[0100]

[0101] The system selected four key events from the event library: E1 (adaptive parameter adjustment), E2 (tool path optimization), E3 (multi-process collaborative control), and E4 (real-time quality feedback control); these events played a significant role in actual processing; for example, when performing P3 (fine grinding), online detection found that the piston hole surface roughness exceeded the standard, immediately triggering the E1 event, the system automatically increased the grinding wheel speed and reduced the feed speed, making the quality return to standard; when performing P5 (fine polishing), the system predicted that there was a risk of tool interference with the complex curved surface, triggering the E2 event in advance, and re-planning a safe and efficient processing path; when P5 was completed, the quality was slightly worse than expected, and when P6 (anodic oxidation) was entered, the E3 event was triggered, automatically extending the oxidation time to compensate for the shortcomings of the previous process; and the E4 event acts like an tireless quality inspector, monitoring the entire process, and responding immediately if any abnormalities are found.

[0102] Reference Figure 5 In step S14, the specific steps are:

[0103] S141: In the multi-process autonomous control event, the polishing data set of the caliper is determined according to the detection of the multi-process autonomous control event, and the primary polishing region of the caliper and the corresponding primary polishing event are determined based on the identification of the polishing data set of the caliper;

[0104] S142: Determine a plurality of sub-abnormal polishing contents based on the identification of the primary polishing event, and determine the abnormal polishing region of the caliper according to the plurality of sub-abnormal polishing contents, the corresponding polishing data and the polishing region corresponding to the caliper;

[0105] S143: Mark the region position and region form of the abnormal polishing region, determine the first re-polishing content according to the region position of the abnormal polishing region and the corresponding polishing wheel, determine the second re-polishing content according to the region form of the abnormal polishing region and the corresponding polishing wheel, and determine the polishing event based on the first re-polishing content, the second re-polishing content and the plurality of polishing images corresponding to the caliper.

[0106] In the embodiments of the present application, various data in the polishing process are collected in real time through the sensor network in the multi-process autonomous control event; the data includes: position coordinates of the polishing tool, speed, feed speed, contact pressure, vibration signal, acoustic signal, etc.; the collected raw data is preprocessed, including filtering, denoising, and outlier rejection; a time series database is established to store dynamic data in the polishing process; key features are extracted through data mining technology to generate structured polishing data sets.

[0107] The clustering algorithm (such as K-means, DBSCAN) is used to divide the polishing data set into regions; the polishing parameter characteristics of each region, such as pressure distribution, speed change, vibration intensity, etc., are analyzed; according to the consistency and stability of the polishing parameters, the primary polishing area that needs to be focused on is identified; the mapping relationship between the polishing area and the polishing parameters is established, and the primary polishing area distribution map is generated, marking the range and characteristics of each region.

[0108] The clustering algorithm (such as K-means, DBSCAN) is used to divide the polishing data set into regions; the polishing parameter characteristics of each region, such as pressure distribution, speed change, vibration intensity, etc., are analyzed; according to the consistency and stability of the polishing parameters, the primary polishing area that needs to be focused on is identified; the mapping relationship between the polishing area and the polishing parameters is established, and the primary polishing area distribution map is generated, marking the range and characteristics of each region.

[0109] Further, based on the identification of the primary polishing event, a plurality of sub-abnormal polishing contents are determined, and the abnormal polishing area of the caliper is determined according to the plurality of sub-abnormal polishing contents, the corresponding polishing data and the polishing area corresponding to the caliper, which comprehensively considers the plurality of sub-abnormal polishing contents, the corresponding polishing data and the polishing area corresponding to the caliper, and ensures the accuracy of the abnormal polishing area of the caliper.

[0110] At this time, the clustering algorithm (such as K-means, DBSCAN) is used to divide the polishing data set into regions; the polishing parameter characteristics of each region, such as pressure distribution, speed change, vibration intensity, etc., are analyzed; according to the consistency and stability of the polishing parameters, the primary polishing area that needs to be focused on is identified; the mapping relationship between the polishing area and the polishing parameters is established, and the primary polishing area distribution map is generated, marking the range and characteristics of each region.

[0111] The sub-abnormal polishing content is associated with the spatio-temporal polishing data to determine the precise position of the abnormality; the spatial statistical analysis method is applied to identify the hot spot area of the abnormality; the importance and influence range of the abnormal area are evaluated in combination with the geometric characteristics and functional requirements of the caliper; the severity of the abnormal area is classified, considering the abnormal intensity, duration and functional impact, and the abnormal polishing area distribution map is generated, marking the position, type, severity and processing priority of each abnormal area.

[0112] Therefore, the region position and region morphology of the abnormal polishing area are marked, the first re-polishing content is determined according to the region position of the abnormal polishing area and the corresponding polishing wheel, the second re-polishing content is determined according to the region morphology of the abnormal polishing area and the corresponding polishing wheel, and the polishing event is determined based on the first re-polishing content, the second re-polishing content and the plurality of polishing images corresponding to the caliper, which comprehensively considers the first re-polishing content, the second re-polishing content and the plurality of polishing images corresponding to the caliper, and ensures the accuracy of the polishing event.

[0113] At this time, the accurate spatial coordinates of the abnormal area are obtained by using a high-precision three-dimensional measurement system (such as a laser scanner or a structured light projector); the surface morphology characteristics of the abnormal area are analyzed by an image processing algorithm, and geometric parameters and surface texture characteristics are extracted; geometric description parameters of the abnormal area are calculated, including area, perimeter, aspect ratio, curvature distribution, roughness, etc.; a position-morphology database of the abnormal area is established to store detailed feature information of each area, and an abnormal area marker map is generated to visually display the spatial distribution and morphological characteristics of each abnormal area on the three-dimensional model.

[0114] The spatial position characteristics of the abnormal area are analyzed to evaluate the accessibility and operating space constraints of the polishing tool; according to the position characteristics, the appropriate polishing wheel type, size and installation method are selected; the feed path of the polishing wheel is planned, including the starting point, ending point, path shape and transition method; the attitude parameters of the polishing wheel are determined, including the inclination angle, contact point and pressure distribution; the first re-polishing strategy is formulated, including the polishing sequence, key areas and parameter settings, and a first re-polishing content document is generated, containing detailed path planning, attitude control and parameter settings.

[0115] The surface morphology characteristics of the abnormal area are analyzed, including curvature variation, texture direction, roughness distribution, etc.; according to the morphological characteristics, the appropriate polishing wheel material, hardness and surface properties are selected; the polishing process parameters are determined, including speed, pressure, time, polishing agent type and concentration, etc.; a polishing strategy is formulated for specific morphological characteristics, such as local reinforcement treatment and texture direction control, and a second re-polishing content document is generated, containing detailed process parameters and quality control points.

[0116] The first and second re-polishing contents are integrated to form a complete polishing plan; similar historical processing cases of the abnormal area are retrieved from the polishing image database for image comparison and effect analysis; the polishing effect is predicted and the polishing parameters are optimized through simulation technology; the optimal polishing event is determined according to the characteristics of the current abnormal area and the historical cases, and a polishing event execution plan is generated, including event trigger conditions, execution steps, quality control standards and emergency plans.

[0117] Reference Figure 6 In step S15, the specific steps are as follows:

[0118] S151: Collect the actual image of the caliper after polishing, determine a plurality of surface abnormal areas according to the actual image, the polishing event and the overall morphology of the caliper, and determine corresponding surface abnormal features based on the identification of the surface abnormal areas to collect a plurality of surface abnormal features;

[0119] S152: Collect the surface processing requirements of the caliper and multiple fine processing tools, determine the first surface autonomous processing coefficient according to the multiple surface abnormal features and the surface processing requirements of the caliper, and determine the second surface autonomous processing coefficient according to the multiple surface abnormal features and the multiple fine processing tools;

[0120] S153: Determine the surface autonomous processing event of the caliper based on the mapping relationship of the first surface autonomous processing coefficient, the second surface autonomous processing coefficient and the surface autonomous processing event; correlate the primary polishing event, the polishing event and the surface autonomous processing event, and construct the multi-process autonomous control system of the caliper according to the multiple training of the primary polishing event, the polishing event and the surface autonomous processing event.

[0121] In the embodiment of the present application, the multi-angle image after polishing of the caliper is collected; the image preprocessing algorithm (such as histogram equalization, Gaussian filtering, non-local mean denoising) is applied to enhance the image quality; the residual abnormal area existing in the image is analyzed in combination with the execution record of the polishing event; the mapping relationship between the three-dimensional coordinate system and the two-dimensional image is established considering the geometric features of the overall morphology of the caliper; the surface abnormal area is identified through the image segmentation algorithm (such as region growing, watershed, deep learning semantic segmentation), and the surface abnormal area distribution map is generated, and the position, size and preliminary classification of each abnormal area are labeled.

[0122] The features of each surface abnormal area are extracted, including geometric features, texture features and optical features; the geometric features include area, perimeter, aspect ratio, circularity, boundary irregularity, etc.; the texture features include gray level co-occurrence matrix features (contrast, energy, entropy, correlation), local binary pattern (LBP) features, Gabor filter response, etc.; the optical features include reflectivity, chroma, brightness distribution, glossiness, etc.; the feature selection algorithm (such as principal component analysis PCA, recursive feature elimination RFE) is applied to screen the most discriminant features; the surface abnormal feature database is established to store the detailed feature information of each abnormal area, and the surface abnormal feature report is generated, including the feature value, the feature distribution and the abnormal type judgment.

[0123] Further, the surface processing requirements of the caliper and multiple fine processing tools are collected, the first surface autonomous processing coefficient is determined according to the multiple surface abnormal features and the surface processing requirements of the caliper, and the second surface autonomous processing coefficient is determined according to the multiple surface abnormal features and the multiple fine processing tools, which is compatible with the overall consideration of the multiple surface abnormal features and the multiple fine processing tools, and ensures the accuracy of the second surface autonomous processing coefficient.

[0124] At this time, feature extraction is performed on each surface anomaly region, including geometric features, texture features and optical features; the geometric features include area, perimeter, aspect ratio, circularity, boundary irregularity, etc.; the texture features include gray level co-occurrence matrix features (contrast, energy, entropy, correlation), local binary pattern (LBP) features, Gabor filter responses, etc.; the optical features include reflectivity, chroma, brightness distribution, glossiness, etc.; a feature selection algorithm (such as principal component analysis PCA, recursive feature elimination RFE) is applied to screen the most discriminative features; a surface anomaly feature database is established to store detailed feature information of each anomaly region, a surface anomaly feature report is generated, including feature values, feature distribution and anomaly type judgment.

[0125] The deviation degree of the surface anomaly features from the surface processing requirements is analyzed, and the deviation index of each feature is calculated; a correlation model of anomaly feature importance and processing requirement strictness is established to determine the weight coefficient of each feature; a multi-attribute decision method (such as TOPSIS, AHP) is applied to calculate the comprehensive deviation degree; the priority and functional influence of the processing requirements are considered to adjust the deviation degree calculation result; the comprehensive deviation degree is normalized to the range of 0-1 to obtain the first surface autonomous processing coefficient, and a coefficient calculation report is generated, including the calculation process, key parameters and result analysis.

[0126] The matching degree of the surface anomaly features and the fine processing tools is analyzed, and the repair ability of each tool to the anomaly features is evaluated; a multi-criteria evaluation model for tool selection is established, considering factors such as repair effect, processing efficiency and tool life; fuzzy comprehensive evaluation or gray correlation analysis is applied to calculate the tool matching degree; the availability, cost and operation complexity of the tool are considered to adjust the matching degree calculation result; the tool matching degree is normalized to the range of 0-1 to obtain the second surface autonomous processing coefficient, and a coefficient calculation report is generated, including the tool evaluation result and the recommended scheme.

[0127] Therefore, the surface autonomous processing event of the caliper is determined based on the mapping relationship between the first surface autonomous processing coefficient, the second surface autonomous processing coefficient and the surface autonomous processing event; the primary polishing event, the polishing event and the surface autonomous processing event are associated, and a multi-process autonomous control system of the caliper is constructed according to the multiple training of the primary polishing event, the polishing event and the surface autonomous processing event, which considers the mapping relationship between the first surface autonomous processing coefficient, the second surface autonomous processing coefficient and the surface autonomous processing event as a whole, ensures the accuracy of the surface autonomous processing event of the caliper, introduces the polishing event, and further autonomously controls the primary polishing event, the polishing event and the surface autonomous processing event, realizing the overall consideration of the primary polishing event, the polishing event and the surface autonomous processing event, and improving the accuracy of the multi-process autonomous control system.

[0128] At this time, a mapping relationship table of surface autonomous machining coefficients and machining events is established, and a machining strategy corresponding to different coefficient combinations is defined;The overall machining demand index is calculated by synthesizing the first and second surface autonomous machining coefficients;According to the overall machining demand index, a suitable machining event is selected from the machining event library;For each selected machining event, define specific trigger conditions, machining parameters and expected effects;Optimize the execution sequence of the machining event, consider the machining efficiency and quality requirements, generate the surface autonomous machining event sequence, including the logical relationship and execution plan between events.

[0129] The time sequence relationship of the three types of events is analyzed, the event chain model is established, the causal relationship and dependency relationship between events are identified, the event association network is constructed, the graph theory algorithm is applied to analyze the critical path and bottleneck node of the event network, the event parameter transmission mechanism is established to ensure that the results of the previous events can be effectively transmitted to the subsequent events, the connection mode between events is optimized to reduce the transition time and resource waste, and the event association graph is generated to intuitively display the overall relationship of the three types of events.

[0130] The time sequence relationship of the three types of events is analyzed, the event chain model is established, the causal relationship and dependency relationship between events are identified, the event association network is constructed, the graph theory algorithm is applied to analyze the critical path and bottleneck node of the event network, the event parameter transmission mechanism is established to ensure that the results of the previous events can be effectively transmitted to the subsequent events, the connection mode between events is optimized to reduce the transition time and resource waste, and the event association graph is generated to intuitively display the overall relationship of the three types of events.

[0131] Please refer to Figure 7 , Figure 7 It is a structure composition schematic view of the multi-process autonomous control system of the factory based on the Internet of Things in the embodiment of the application;The multi-process autonomous control system of the factory based on the Internet of Things comprises:

[0132] The complete curved surface feature module 21 is used for multi-process machining of the caliper in the factory, and the surface body machining image of the caliper is collected based on the Internet of Things and the camera, and each complete curved surface feature is determined according to the recognition of the surface body machining image;

[0133] The key machining node module 22 is used for determining the corresponding curved surface machining route according to the feature position of each complete curved surface feature and the overall form of the caliper;A plurality of key machining nodes are determined according to the curved surface machining route and the current posture of the caliper;

[0134] The multi-process autonomous control event module 23 is used for determining a multi-level surface treatment system according to the plurality of key machining nodes and the process table of the caliper, determining a plurality of surface treatment items based on the recognition of the multi-level surface treatment system, and determining the multi-process autonomous control event of the caliper in the surface treatment stage according to the item content of the plurality of surface treatment items, the corresponding surface treatment part and the overall form of the caliper;

[0135] The polishing event module 24 is configured to determine a primary grinding event based on the multi-process autonomous control event; determine an abnormal grinding area of the caliper according to the identification of the primary grinding event, and determine a polishing event according to the abnormal grinding area and a subsequent polishing process;

[0136] The multi-process autonomous control system module 25 is configured to determine a plurality of surface abnormal features according to the polishing event and the corresponding overall morphology, determine a surface autonomous processing event of the caliper according to the plurality of surface abnormal features, surface processing requirements of the caliper and a plurality of fine processing tools, and construct a multi-process autonomous control system of the caliper based on the primary grinding event, the polishing event and the surface autonomous processing event.

[0137] Any combination of the technical features of the above embodiments is possible. In order to make the description concise, not all combinations of the technical features in the above embodiments are described, however, as long as the combinations of the technical features do not exist, they should be considered as the scope of the present disclosure.

Claims

1. A multi-process autonomous control method for an Internet of Things-based factory that processes calipers, characterized by, The caliper is processed in multiple processes in a factory, surface body processing images of the caliper are collected based on Internet of Things and cameras, and each complete curved surface feature is determined according to recognition of the surface body processing images; A corresponding curved surface processing route is determined according to a feature position of each complete curved surface feature and an overall shape of the caliper, and a plurality of key processing nodes are determined according to the curved surface processing route and a current posture of the caliper; A multi-level surface treatment system is determined according to the plurality of key processing nodes and a process table of the caliper, a plurality of surface treatment items are determined based on recognition of the multi-level surface treatment system, and a plurality of self-control events of the caliper in a surface treatment stage are determined according to item content of the plurality of surface treatment items, corresponding surface treatment parts and the overall shape of the caliper; A primary polishing event is determined based on the plurality of self-control events, an abnormal polishing area of the caliper is determined according to recognition of the primary polishing event, and a polishing event is determined according to the abnormal polishing area and a subsequent polishing process; In the plurality of self-control events, a polishing data set of the caliper is determined according to detection of the plurality of self-control events, a primary polishing area of the caliper and a corresponding primary polishing event are determined based on recognition of the polishing data set of the caliper; A plurality of sub-abnormal polishing contents are determined based on recognition of the primary polishing event, an abnormal polishing area of the caliper is determined according to the plurality of sub-abnormal polishing contents, corresponding polishing data and a polishing area corresponding to the caliper; A region position and a region shape of the abnormal polishing area are marked, a first re-polishing content is determined according to the region position of the abnormal polishing area and a corresponding polishing wheel, a second re-polishing content is determined according to the region shape of the abnormal polishing area and the corresponding polishing wheel, and the polishing event is determined based on the first re-polishing content, the second re-polishing content and a plurality of polishing images corresponding to the caliper; A plurality of surface abnormal features are determined according to the polishing event and the corresponding overall shape, a surface self-processing event of the caliper is determined according to the plurality of surface abnormal features, surface processing requirements of the caliper and a plurality of fine processing tools, and a multi-process self-control system of the caliper is constructed based on the primary polishing event, the polishing event and the surface self-processing event. The caliper is processed in multiple processes in a factory, surface body processing images of the caliper are collected based on Internet of Things and cameras, and each complete curved surface feature is determined according to recognition of the surface body processing images; 2. The multi-process autonomous control method of an Internet of Things-based factory of machining calipers according to claim 1, characterized in that, The caliper is processed in multiple processes in a factory, surface body processing images of the caliper are collected based on Internet of Things and cameras, and each complete curved surface feature is determined according to recognition of the surface body processing images; The caliper is processed in multiple processes in a factory, surface body processing images of the caliper are collected based on Internet of Things and cameras, and each complete curved surface feature is determined according to recognition of the surface body processing images; ​ 3. The multi-process autonomous control method of an Internet of Things-based factory of machining calipers according to claim 1, characterized in that, ​ Collecting each complete curved surface feature, determining the feature position of each complete curved surface feature based on the detection of each complete curved surface feature, collecting the overall shape of the caliper and the corresponding surface processing requirements, and determining the curved surface processing route of the caliper according to the feature position of the plurality of complete curved surface features, the overall shape of the caliper and the corresponding surface processing requirements; Real-time monitoring of the caliper and collection of a plurality of posture parameters of the caliper, determination of the current posture of the caliper according to the plurality of posture parameters of the caliper, and determination of the corresponding curved surface processing tool according to the current posture of the caliper, the corresponding curved surface processing content and the processing tool database; and determination of a plurality of key processing nodes of the caliper based on the curved surface processing route of the caliper, the current posture of the caliper and the corresponding curved surface processing tool.

4. The multi-process autonomous control method of an Internet of Things-based factory of machining calipers according to claim 1, characterized in that, According to the plurality of key processing nodes and the process table of the caliper, a multi-level surface treatment system is determined, a plurality of surface treatment items are determined based on the identification of the multi-level surface treatment system, and a plurality of process self-control events of the caliper in the surface treatment stage are determined according to the item content of the plurality of surface treatment items, the corresponding surface treatment part and the overall shape of the caliper, including: Based on the detection of the plurality of key processing nodes, the node position and the corresponding node shape of the plurality of key processing nodes are determined, and the process table of the caliper is determined based on the model, the overall shape and the factory processing database of the caliper; and the multi-level surface treatment system is determined according to the node position of the plurality of key processing nodes, the corresponding node shape and the process table of the caliper. According to the identification of the multi-level surface treatment system, a plurality of surface treatment items are determined, and the item content of the plurality of surface treatment items is marked, and the corresponding surface treatment part is determined according to the item content of the plurality of surface treatment items and the current posture of the caliper.

5. The multi-process autonomous control method of an Internet of Things based factory of machining calipers according to claim 4, characterized in that, According to the plurality of key processing nodes and the process table of the caliper, a multi-level surface treatment system is determined, a plurality of surface treatment items are determined based on the identification of the multi-level surface treatment system, and a plurality of process self-control events of the caliper in the surface treatment stage are determined according to the item content of the plurality of surface treatment items, the corresponding surface treatment part and the overall shape of the caliper, further including: According to the item content of the plurality of surface treatment items and the overall shape of the caliper, a first self-control coefficient is determined, a second self-control coefficient is determined according to the surface treatment part corresponding to the plurality of surface treatment items and the overall shape of the caliper, and the plurality of process self-control events of the caliper in the surface treatment stage are determined based on the first self-control coefficient, the second self-control coefficient and the self-control event mapping relationship of the surface treatment stage.

6. The multi-process autonomous control method of an Internet of Things-based factory of machining calipers according to claim 1, characterized in that, According to the polishing event and the corresponding overall shape, a plurality of surface abnormal features are determined, and the surface self-processing event of the caliper is determined according to the plurality of surface abnormal features, the surface processing requirements of the caliper and a plurality of fine processing tools; Based on the primary polishing event, the polishing event and the surface self-processing event, a multi-process self-control system of the caliper is constructed, including: Collecting the actual image of the caliper after polishing, determining a plurality of surface abnormal regions according to the actual image, the polishing event and the overall shape of the caliper, and determining the corresponding surface abnormal feature based on the identification of the surface abnormal region in the plurality of surface abnormal regions to collect a plurality of surface abnormal features.

7. The multi-process autonomous control method of an Internet of Things-based factory of machining calipers according to claim 6, characterized in that, The multiple surface abnormal features are determined according to the polishing event and the corresponding overall morphology, the surface autonomous processing event of the caliper is determined according to the multiple surface abnormal features, the surface processing requirements of the caliper and the multiple fine processing tools; The multiple-process autonomous control system of the caliper is constructed based on the primary polishing event, the polishing event and the surface autonomous processing event, and further comprises: The surface processing requirements of the caliper and the multiple fine processing tools are collected, the first surface autonomous processing coefficient is determined according to the multiple surface abnormal features and the surface processing requirements of the caliper, and the second surface autonomous processing coefficient is determined according to the multiple surface abnormal features and the multiple fine processing tools; The surface autonomous processing event of the caliper is determined based on the mapping relationship of the first surface autonomous processing coefficient, the second surface autonomous processing coefficient and the surface autonomous processing event; the primary polishing event, the polishing event and the surface autonomous processing event are associated, and the multiple-process autonomous control system of the caliper is constructed according to the multiple training of the primary polishing event, the polishing event and the surface autonomous processing event.

8. A multi-process autonomous control system of an Internet of Things-based factory, characterized by, The multiple-process autonomous control system of the factory based on the Internet of Things is applied to the multiple-process autonomous control method of the caliper processing factory based on the Internet of Things as claimed in any one of claims 1-7, and comprises: A complete curved surface feature module is used for multiple-process processing of the caliper in the factory, and the surface body processing image of the caliper is collected based on the Internet of Things and the camera, and each complete curved surface feature is determined according to the recognition of the surface body processing image; A key processing node module is used for determining the corresponding curved surface processing route according to the feature position of each complete curved surface feature and the overall morphology of the caliper, and determining multiple key processing nodes according to the curved surface processing route and the current posture of the caliper; A multiple-process autonomous control event module is used for determining a multiple-level surface treatment system according to the multiple key processing nodes and the process table of the caliper, determining multiple surface treatment items based on the recognition of the multiple-level surface treatment system, and determining the multiple-process autonomous control event of the caliper in the surface treatment stage according to the item content of the multiple surface treatment items, the corresponding surface treatment part and the overall morphology of the caliper; A polishing event module is used for determining the primary polishing event based on the multiple-process autonomous control event, determining the abnormal polishing area of the caliper according to the recognition of the primary polishing event, and determining the polishing event according to the abnormal polishing area and the subsequent polishing process; A multiple-process autonomous control system module is used for determining multiple surface abnormal features according to the polishing event and the corresponding overall morphology, determining the surface autonomous processing event of the caliper according to the multiple surface abnormal features, the surface processing requirements of the caliper and the multiple fine processing tools, and constructing the multiple-process autonomous control system of the caliper based on the primary polishing event, the polishing event and the surface autonomous processing event.

Images