Footwear processing method based on three-dimensional model and related equipment

Abstract

The invention relates to the technical field of intelligent manufacturing, in particular to a shoe processing method based on a three-dimensional model and related equipment. A footwear processing method based on a three-dimensional model comprises the steps that footwear is scanned in an omnibearing mode through a 3D visual sensor to obtain three-dimensional data, an actual footwear model is constructed, a five-axis robot carries out precise processing according to the model, and processing deviation is monitored in real time. According to the method, the shoe processing precision and efficiency are improved, and human intervention and production cost are reduced.

Description

Technical Field

[0001] The present application relates to the field of intelligent manufacturing technology, and in particular to a footwear processing method based on a three-dimensional model and related equipment. Background Art

[0002] In the footwear manufacturing industry, traditional processing methods often rely on manual operations or simple mechanized production, which is not only inefficient but also difficult to ensure processing accuracy. With the development of science and technology, especially the advancement of 3D scanning and robotics, new possibilities have been brought to the footwear processing industry. However, existing automated processing solutions still face problems such as inaccurate data acquisition, unreasonable processing path planning, and inability to monitor processing deviations in real time.

[0003] Therefore, based on the above problems, the prior art still needs to be improved. Summary of the invention

[0004] The purpose of this application is to provide a shoe processing method and related equipment based on a three-dimensional model, aiming to solve the problem of insufficient precision and efficiency in shoe processing.

[0005] The first object of this application is to provide a footwear processing method based on a three-dimensional model, comprising: Acquire the material and shape of the footwear, and pre-set the scanning frequency and range of the 3D vision sensor based on the material and shape; Performing a full-scale scan of the shoes on the production line to obtain three-dimensional data of each part of the shoes, and monitoring and compensating for data errors caused by material deformation in real time during the scanning process; Using an image processing and modeling algorithm that is adapted to the characteristics of the shoe data and optimized, preprocessing and feature extraction are performed on the three-dimensional data, and the three-dimensional data is compared and verified with a standard shoe model to construct an actual shoe model of the shoe; The constructed actual shoe model and related parameters and instructions are transmitted to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; The five-axis robot control system receives and analyzes the data instructions, and plans the precise processing path and action sequence according to the shoe shape, size, and material distribution information contained in the actual shoe model and in combination with the pre-optimized processing parameter library; Based on the motion accuracy limitation and possible interference of the five-axis robot, the processing path and the action sequence are adjusted and optimized in real time; The five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by tool wear and material property change is monitored and obtained in real time.

[0006] By adopting the above technical solution, by presetting the scanning frequency and range of the 3D vision sensor, accurate scanning can be achieved according to the material and shape of the footwear, improving the pertinence and efficiency of the scanning. All-round scanning and real-time monitoring and compensation of data errors ensure the accuracy and completeness of the acquired three-dimensional data, providing a reliable data basis for subsequent processing. The actual footwear model is constructed using an optimized algorithm, and is transmitted to the five-axis robot control system in real time through a data transmission channel, realizing the rapid and accurate transmission of data and improving the timeliness and coordination of the entire processing process. The five-axis robot control system can plan accurate processing paths and action sequences based on the actual footwear model and the optimized processing parameter library, improving the accuracy and quality of processing. Real-time adjustment and optimization based on the motion accuracy limit and possible interference of the five-axis robot further ensure the stability and safety of the processing and reduce the risk of processing errors. During the processing process, real-time monitoring and acquisition of processing deviations caused by various factors can timely discover problems and take corresponding measures to adjust, ensuring the quality consistency and reliability of the final footwear products.

[0007] In a possible implementation manner of the present application, the five-axis robot processes the footwear according to the processing path and action sequence. During the processing, the step of real-time monitoring and obtaining the processing deviation caused by tool wear and material property change factors includes: Acquiring real-time status data of the footwear when it is processed by monitoring with a variety of sensors configured on the five-axis robot; Comparing the state data with preset standard data or models and feeding back the results to obtain test results, which include analyzing the consistency between the actual material removal amount and the expected amount, the deviation between the processed shape and the actual shoe model, the degree of tool wear and the change in material properties; Based on the test results, re-evaluate the processing of the footwear and obtain an evaluation result; Update the actual shoe model and determine a new processing strategy according to the evaluation result, and dynamically adjust the processing path and the action sequence; The above-mentioned cycle of monitoring, feedback, evaluation and adjustment is continuously carried out in subsequent processing to obtain and deal with processing deviations caused by various factors in real time.

[0008] By adopting the above technical solutions and obtaining real-time status data through monitoring with multiple sensors, the real-time status of footwear processing can be fully and accurately grasped without missing any subtle changes. The status data is compared with the preset standards and the test results are fed back, making the control of the processing process more rigorous and scientific, and timely discovering the differences between the actual material removal amount, processing shape and other aspects and the expected. Re-evaluating the processing situation based on the test results can quickly and accurately judge the pros and cons of the current processing, providing a reliable basis for subsequent adjustments. According to the evaluation results, the model is updated and new strategies are determined, and the processing path and action sequence are dynamically adjusted to achieve flexible response of the processing process, minimize the adverse effects caused by factors such as tool wear and changes in material properties, and ensure the accuracy and quality of processing. Continuously circulating the above process in subsequent processing can continuously optimize the processing effect, ensure that each pair of shoes can meet high quality standards, and improve the consistency and stability of the product.

[0009] In a possible implementation of the present application, the method includes: The end effector of the five-axis robot and the shoe part being processed are scanned and monitored in real time by the 3D vision sensor to obtain actual processing position information; Comparing and analyzing the actual processing position information with the preset processing position to determine whether the processing position of the five-axis robot is accurate; If the processing position of the five-axis robot is inaccurate, the deviation information is promptly fed back to the control system of the five-axis robot through the 3D vision sensor, and the control system then adjusts the processing path and action sequence of the five-axis robot to correct the deviation; If the processing position of the five-axis robot is accurate, the five-axis robot continues to perform processing operations according to the predetermined processing path and action sequence.

[0010] By adopting the above technical solution, the end effector and processing parts of the five-axis robot are scanned and monitored in real time using 3D vision sensors, which can accurately obtain the actual processing position information and ensure the control of the details of the processing process. By comparing and analyzing the actual processing position with the preset position, position deviations can be discovered in a timely manner, improving the accuracy and reliability of processing. If the position is inaccurate, the deviation information can be quickly fed back to the control system and adjusted, effectively avoiding processing errors and product quality problems caused by position deviations, and reducing the scrap rate. When the position is accurate, the five-axis robot continues processing according to the established path and sequence, ensuring the efficiency and stability of processing, reducing unnecessary interruptions and adjustments, and improving production efficiency.

[0011] In a possible implementation of the present application, the scanning frequency and range are pre-set according to the material and shape of the footwear, and the steps of performing an all-round scanning on the footwear on the production line include: Obtaining and analyzing the material and shape of the footwear; Based on the material and shape, determining the material properties and shape characteristics of the footwear; Determining a scanning frequency of the 3D vision sensor based on reflectivity and transparency of the material properties and complexity of the shape characteristics; Acquiring shape characteristics and preset key part information of the footwear; Determine the scanning range of the 3D vision sensor so that the scanning range can completely cover all parts of the shoe, and at the same time perform a more concentrated and detailed scanning on the key parts corresponding to the key part information; The 3D vision sensor is activated to perform an all-round scan of the shoes continuously moving on the production line. During the scanning process, the quality and integrity of the scanned data are monitored in real time.

[0012] By adopting the above technical solutions, the material and shape of the footwear are obtained and analyzed, laying the foundation for subsequent precise scanning, making the entire scanning process more targeted and scientific. Determining the characteristics and features based on the material and shape can provide a deeper understanding of the characteristics of the footwear, thereby setting the scanning parameters more accurately. Determining the scanning frequency based on the reflectivity, transparency and complexity of the shape in the material characteristics can avoid inaccurate or missing data caused by improper scanning frequency and improve the quality of the scanned data. Obtaining shape characteristics and key parts information and determining the scanning range can not only ensure full coverage, but also focus on scanning key parts, improving the efficiency and effectiveness of the scan. Starting the 3D visual sensor for all-round scanning and real-time monitoring of data quality and integrity can promptly discover and solve problems in the scanning process, ensure the acquisition of high-quality and complete scanning data, and provide a reliable foundation for subsequent processing and handling.

[0013] In a possible implementation manner of the present application, the method further includes: After the processing of the footwear is completed, the 3D vision sensor is started to perform an all-round scan of the footwear to detect defects on the surface and inside of the footwear, and obtain defect detection results; Generate defect information according to the defect detection result, wherein the defect information includes the type, location, size and shape of the defect; The defect information is sent to the control system of the five-axis robot through a data transmission channel; after receiving the defect information, the five-axis robot control system plans a repair action and path according to the characteristics and location of the defect; After the repair is completed, the repaired part is scanned again by the 3D vision sensor to evaluate the repair effect.

[0014] If the evaluated repair effect does not meet the requirements, repeat the above steps to perform repair again, and the evaluated repair effect meets the requirements.

[0015] By adopting the above technical solution, the full-scale scanning and defect detection after processing can ensure timely detection of problems on the surface and inside of the footwear, without missing any minor defects, and improving the quality control ability of the product. Generate detailed defect information, including type, location, size and shape, etc., to provide accurate guidance for subsequent repair work, improve the pertinence and efficiency of the repair. Send the defect information to the control system and plan the repair action and path, realize automated repair planning, and reduce the time and error of manual judgment and decision-making. After the repair, scan again to evaluate the repair effect, which can effectively supervise and inspect the repair work and ensure that the repair quality meets the requirements. If the evaluation does not meet the requirements, repeat the repair until it meets the requirements. This strict quality control mechanism can guarantee the quality stability and reliability of the product to the greatest extent.

[0016] In a possible implementation of the present application, if the evaluation and repair effect does not meet the requirements, the above steps are repeated to perform repair again, and the steps after the evaluation and repair effect meets the requirements include: Collect the shoe data after repair, including the original data with defects, the repair process data and the final data after repair, and mark the defect type, location and severity in these data in detail; Using the labeled data to train a deep learning model, so that the deep learning can learn the relationship between the characteristics of the defect and the repair method and effect; After the repair is completed, the footwear is scanned again by the 3D vision sensor to obtain detailed information after the repair; Extracting defect-related features from the repaired details; The extracted features are input into the trained deep learning model to predict the cause of the defect and the factors in the processing link that may cause it, and obtain the prediction results; Generate a detailed feedback report based on the prediction results, the feedback including possible causes of defects, related processing parameters and actions; Transmitting the feedback report to the control system of the five-axis robot; After the five-axis robot adjusts the processing parameters and processing actions according to the feedback report, new processing result data is obtained, and the new processing result data is collected again and used to update the deep learning model.

[0017] By adopting the above technical solutions, the data related to the repair is collected and annotated, which provides rich and accurate materials for subsequent analysis and learning, making the understanding of the entire repair process more in-depth and comprehensive. Using annotated data to train the deep learning model helps the model learn the complex relationship between defects, repair methods and effects, thereby improving the ability to predict the causes of defects and problems in the processing link. Scanning again to obtain detailed information after the repair and extract features can provide more accurate input for the model's prediction and enhance the accuracy of the prediction. By predicting the causes of defects and possible processing factors, a detailed feedback report is generated, which provides a clear direction and basis for subsequent improvements, making the problem solving more targeted and effective. The feedback report is transmitted to the control system to adjust the processing parameters and actions, which can timely optimize the processing process, reduce the probability of defects, and improve the overall processing quality and efficiency. Continuously collecting new processing result data to update the deep learning model, the model is continuously optimized and improved to adapt to different processing conditions and problems, and further improve the optimization ability of the footwear processing process.

[0018] In a possible implementation of the present application, based on the motion accuracy limitation and possible interference of the five-axis robot, the step of adjusting and optimizing the processing path and the action sequence in real time includes: Acquire the current motion accuracy data of the five-axis robot in real time, including position accuracy and posture accuracy, and compare them with the preset accuracy requirements; Monitoring the spatial positional relationship between the components of the five-axis robot, the tool, and the footwear being processed, and predicting whether there is a risk of interference through simulation and calculation; Conduct a detailed analysis of the currently planned machining path to evaluate the movement difficulty and impact on accuracy of each point on the machining path; Check the starting and ending states, speed changes, and acceleration changes of each action in the action sequence, and analyze the smoothness of the connection between each action; Determine the adjustment strategy based on the accuracy and interference detection results. If the accuracy does not meet the standard or there is a risk of interference, modify the processing path and optimize the action sequence. If the accuracy meets the standard and there is no risk of interference but there is room for optimization, further fine-tune the processing path and the action sequence. regenerating the processing path and the action sequence according to the determined adjustment strategy; The re-planned processing path and the action sequence are quickly simulated and verified, and applied to actual processing in real time after passing the verification.

[0019] By adopting the above technical solutions, the motion accuracy data of the five-axis robot can be obtained in real time and compared with the preset requirements, so that the accuracy deviation can be discovered in time, the accuracy of processing can be ensured, and the product quality can be improved. Monitoring the spatial position relationship between parts, tools and footwear to predict the risk of interference can effectively avoid collision and damage, reduce equipment failure and maintenance costs. Detailed analysis of the processing path and evaluation of the difficulty of movement and the impact on accuracy can help plan a more reasonable and efficient processing path and improve processing efficiency. Checking the relevant parameters of each action in the action sequence and analyzing the smoothness of the action connection can make the processing process smoother, reduce vibration and impact, and extend the service life of the equipment. Determining the adjustment strategy based on the test results can flexibly respond to various situations. Modification and optimization when the accuracy is not up to standard or there is a risk of interference ensures the safety and accuracy of processing; fine-tuning when the accuracy is up to standard but there is room for optimization further improves the processing effect and efficiency. Regenerate the path and action sequence according to the adjustment strategy, and conduct rapid simulation verification to ensure the effectiveness and feasibility of the adjustment and reduce the risk of errors in actual processing. After verification, it is applied to actual processing in real time, which can quickly achieve optimization effects and enhance the timeliness and adaptability of production.

[0020] The second object of the present application is to provide a footwear processing system based on a three-dimensional model, the system comprising: Scanning frequency and range preset module: obtains the material and shape of the footwear, and pre-sets the scanning frequency and range of the 3D vision sensor based on the material and shape; 3D data acquisition module: performs all-round scanning of the shoes on the production line to obtain 3D data of each part of the shoes, and monitors and compensates for data errors caused by material deformation in real time during the scanning process; Actual shoe model construction module: using image processing and modeling algorithms adapted to the characteristics of the shoe data and optimized to pre-process and extract features from the three-dimensional data, comparing and verifying the three-dimensional data with a standard shoe model, and constructing an actual shoe model of the shoe; Actual shoe model transmission module: transmits the constructed actual shoe model and related parameters and instructions to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; Processing path and action sequence planning module: The five-axis robot control system receives and analyzes data instructions, and plans precise processing paths and action sequences according to the shoe shape, size, and material distribution information contained in the actual shoe model and in combination with a pre-optimized processing parameter library; Processing path and action sequence optimization module: based on the motion accuracy limit and possible interference of the five-axis robot, the processing path and the action sequence are adjusted and optimized in real time; Processing deviation real-time monitoring module: the five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by tool wear and material property changes is monitored and obtained in real time.

[0021] By adopting the above technical solution, by presetting the scanning frequency and range of the 3D vision sensor, accurate scanning can be achieved according to the material and shape of the footwear, improving the pertinence and efficiency of the scanning. All-round scanning and real-time monitoring and compensation of data errors ensure the accuracy and completeness of the acquired three-dimensional data, providing a reliable data basis for subsequent processing. The actual footwear model is constructed using an optimized algorithm, and is transmitted to the five-axis robot control system in real time through a data transmission channel, realizing the rapid and accurate transmission of data and improving the timeliness and coordination of the entire processing process. The five-axis robot control system can plan accurate processing paths and action sequences based on the actual footwear model and the optimized processing parameter library, improving the accuracy and quality of processing. Real-time adjustment and optimization based on the motion accuracy limit and possible interference of the five-axis robot further ensure the stability and safety of the processing and reduce the risk of processing errors. During the processing process, real-time monitoring and acquisition of processing deviations caused by various factors can timely discover problems and take corresponding measures to adjust, ensuring the quality consistency and reliability of the final footwear products.

[0022] The third object of this application is to provide a shoe processing device based on a three-dimensional model, the device comprising: A memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the above-mentioned shoe processing method based on a three-dimensional model.

[0023] The fourth objective of this application is to provide a storage medium.

[0024] The fourth application objective of this application is achieved through the following technical solutions: A storage medium stores a computer program that can be loaded by a processor and execute the above-mentioned shoe processing method based on a three-dimensional model.

[0025] In summary, the present application includes at least one of the following beneficial technical effects: 1. The 3D vision sensor is used for all-round scanning, which can accurately obtain the three-dimensional data of the footwear, thereby improving the processing accuracy.

[0026] 2. The five-axis robot plans the processing path according to the actual shoe model and can intelligently adapt to the shapes and sizes of different shoes, improving the flexibility and efficiency of processing.

[0027] 3. By real-time monitoring and obtaining processing deviations, problems can be discovered and adjusted in a timely manner, reducing scrap rates and saving production costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic flow chart of a footwear processing method based on a three-dimensional model provided in an embodiment of the present application; Figure 2 It is a virtual structural schematic diagram of a footwear processing system based on a three-dimensional model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0030] In addition, the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article, unless otherwise specified, generally means that the associated objects before and after are in an "or" relationship.

[0031] The embodiments of the present application are further described in detail below in conjunction with the drawings in the specification.

[0032] The present application embodiment provides a shoe processing method based on a three-dimensional model, referring to Figure 1 , the main process of the method is described as follows: S1: Acquire the material and shape of the footwear, and pre-set the scanning frequency and range of the 3D vision sensor based on the material and shape; Among them, before the shoes enter the production process, the material and shape information of the shoes are first obtained through manual input or automated testing equipment. For the judgment of the material, it may be necessary to use material testing instruments to analyze its composition to determine whether it is leather, cloth, plastic or other special materials. At the same time, the shape of the shoes is obtained through image recognition technology or measuring tools, including specific parameters such as the length, width, height, and curvature of the shoe. Based on the acquired material and shape information, the system sets the scanning frequency and range of the 3D vision sensor according to preset rules. If it is a highly reflective material, such as bright leather, the scanning frequency will be appropriately increased to obtain clearer and more accurate data; for materials with higher transparency, such as certain special plastics, the scanning range may focus on the detection of internal structure and thickness. For shoes with more complex shapes, such as high-heeled shoes with complex decorations, the scanning range will cover all decorative details and irregular parts; while for shoes with simple shapes, such as flat cloth shoes, the scanning range is mainly concentrated on key parts, such as the connection between the sole and the upper. When setting the scanning frequency and range, historical data and empirical models are also referenced to ensure that the set parameters can meet the scanning requirements of shoes of different materials and shapes, thereby obtaining high-quality three-dimensional data.

[0033] S2: Performing a full-scale scan of the shoes on the production line to obtain three-dimensional data of various parts of the shoes, and monitoring and compensating for data errors caused by material deformation in real time during the scanning process; Among them, multiple high-precision 3D vision sensors are installed at key positions of the footwear production line to form an all-round scanning network. When the footwear products move at a constant speed on the conveyor belt, these sensors start working at the same time. The sensors use structured light scanning technology or laser scanning technology to quickly project light to various parts of the footwear and receive the reflected light to obtain the initial three-dimensional data. While scanning, strain sensors are installed at key parts of the footwear to monitor the deformation of the material in real time. These sensors transmit the deformation data to the data processing system in real time. After receiving the original three-dimensional data and deformation data, the data processing system immediately analyzes it. For common material deformation modes, such as stretching, compression, bending, etc., the system uses pre-trained mathematical models and algorithms to calculate the impact of deformation on the data. If it is found that the data of a certain part has errors due to deformation, the system will compensate and correct the error data through interpolation, fitting and other methods based on the deformation and the physical properties of the material. For example, when the leather of the upper is stretched during processing, causing the scanned data to show that the length of the upper is too long, the system will calculate the appropriate correction value based on the elastic coefficient and stretching degree of the leather, and adjust the data to obtain accurate three-dimensional data. Throughout the entire process, scanning, monitoring, calculation and compensation are all carried out in real time and continuously, ensuring that the three-dimensional data of each part of the shoe can truly reflect the actual shape of the shoe and is not affected by material deformation.

[0034] S3: using an image processing and modeling algorithm adapted to the characteristics of the shoe data and optimized to preprocess and extract features from the three-dimensional data, comparing and verifying the three-dimensional data with a standard shoe model, and constructing an actual shoe model of the shoe; After obtaining the three-dimensional data of the footwear, it is preprocessed using image processing software and algorithms specially customized for this type of shoe data. The noise and outliers in the data are removed and smoothed by filtering algorithms. In the feature extraction stage, specific algorithms are used according to the type of footwear (such as sports shoes, leather shoes, sandals, etc.) and its unique data characteristics. For example, for sports shoes, the focus is on extracting the texture characteristics of the sole and the distribution characteristics of the air holes on the upper; for leather shoes, the focus is on extracting the cortical texture characteristics of the upper and the shape characteristics of the heel. Then, the processed three-dimensional data is compared and verified with the pre-established standard footwear model. The standard model covers the ideal form and size parameters of various common footwear. When comparing, the size differences between the two in key parts (such as toe shape, heel height, sole thickness, etc.) and the similarity of the overall shape are calculated. Based on the comparison results, the three-dimensional data is adjusted and optimized to construct a footwear model that conforms to the actual situation. In the modeling process, surface reconstruction algorithms, parametric modeling techniques, etc. may be used to ensure the accuracy and detail precision of the model. For example, for a newly designed women's sandal, the weaving texture features of the shoelaces and the unique shape features of the heel are extracted. After comparing with the standard sandal model, it is found that the heel height is slightly higher than the standard. By adjusting the model parameters, an accurate model that matches the actual sandal is constructed.

[0035] S4: The constructed actual shoe model and related parameters and instructions are transmitted to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; Among them, after the actual footwear model is constructed and the relevant parameters and instructions are determined, the data is encoded into a specific format so that it can be accurately identified and parsed during the transmission process. The selected data transmission channel has multiple redundant lines. For example, in addition to the main transmission line, there is also a backup line. When the main line fails or the transmission is blocked, the system can automatically switch to the backup line to ensure that the data transmission is not interrupted. In the data transmission process, error correction coding technology is used. Each piece of data contains additional verification information. After receiving the data, the receiving end will check the integrity and accuracy of the data based on the verification information. If the data is found to be incorrect, a request will be sent to the sending end to request the data to be resent until the correct data is received. In order to achieve real-time transmission, the data transmission channel adopts a high-speed communication protocol, such as the real-time extension protocol of Ethernet or a dedicated industrial real-time communication protocol, to ensure low latency and high bandwidth. At the same time, a data buffer area is set at the key nodes of the transmission. When the data transmission speed fluctuates briefly, the buffer area can temporarily store the data to avoid data loss, and continue to send after the transmission returns to normal. For example, when the actual shoe model and related parameters and instructions are ready to be transmitted to the control system of the five-axis robot, the system first transmits them quickly through the main line. If the main line is suddenly disturbed and some data is wrong, the system immediately switches to the backup line and retransmits and corrects the wrong data to ensure that all data reaches the control system of the five-axis robot accurately and in real time.

[0036] S5: The five-axis robot control system receives and parses the data instructions, and plans a precise processing path and action sequence according to the shoe shape, size, and material distribution information contained in the actual shoe model and in combination with a pre-optimized processing parameter library; Among them, when the control system of the five-axis robot receives the actual shoe model and related data instructions from the data transmission channel, it first decodes and analyzes these data. The system will extract key information such as shoe shape (such as the curvature of the toe, the shape of the heel), size (length, width, height, etc.) and material distribution (material type and thickness of different parts) contained in the actual shoe model. At the same time, the system will access the pre-optimized processing parameter library. This parameter library stores parameters such as the optimal processing speed, tool selection, cutting depth, etc. for various shoe materials and shapes. Based on the extracted shoe model information and the data in the parameter library, the control system uses a path planning algorithm to plan a precise processing path. For example, for the stitching line of the upper, the optimal trajectory of the robot arm movement is planned to ensure that the stitching line is straight and uniform. For the planning of the action sequence, the system will take into account the order of the processing process, the continuity of the action, and the avoidance of unnecessary pauses and repeated actions. For example, the sole is polished first, and then the upper is decorated, and the transition between each action is smooth and natural. During the planning process, the processing path and action sequence are simulated in real time, and virtual simulation is used to check whether there is a risk of collision or an unreachable processing position. If a problem is found, the path and action sequence are adjusted in time until the requirements are met. For example, when processing a pair of sports shoes, the control system selects the appropriate tool and processing speed from the parameter library according to the shape and size of the shoe, and plans a continuous processing path from the sole to the upper, as well as the accurate action sequence of the robot arm, including the tool rotation angle, the timing of feeding and retracting, etc., to ensure efficient and accurate completion of the processing task.

[0037] S6: Based on the motion accuracy limitation and possible interference of the five-axis robot, the processing path and the action sequence are adjusted and optimized in real time; Among them, during the process of footwear processing by the five-axis robot, the control system will continuously monitor the robot's motion state in real time. The high-precision encoders and sensors installed at the robot's joints are used to obtain its motion accuracy data such as position, speed, acceleration, etc. At the same time, the system will use the collision detection algorithm to simulate the spatial position relationship between the robot's various components, tools, and the footwear being processed in real time. If possible interference is predicted, such as the robot arm is about to collide with a protruding part of the footwear, or the tool is too close to other parts, the system will immediately issue an alarm. For the detected motion accuracy deviation, such as the deviation between the actual position and the planned path exceeds the allowable error range, the system will calculate the amount to be adjusted according to the size and direction of the deviation. When adjusting and optimizing the processing path, the system may recalculate the coordinates of the path points and increase or decrease the number of path points to achieve smoother and more accurate movement. For the action sequence, the start time, duration, speed change curve, etc. of the action may be adjusted to adapt to the accuracy limit and avoid interference. For example, if it is found that the robot's accuracy decreases due to excessive movement speed when processing the heel part, the system will reduce the movement speed of the path section and adjust the time interval of subsequent actions accordingly to ensure the smoothness and accuracy of the entire processing process. Or, if the robot is detected to be turning and may interfere with the decoration on the upper of the shoe, the system will immediately modify the path to allow the robot arm to bypass the obstacle, while optimizing the action sequence to avoid unnecessary actions and improve processing efficiency. The entire adjustment and optimization process is completed within milliseconds, ensuring that the five-axis robot can always process shoes in a safe and precise state.

[0038] S7: The five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by factors such as tool wear and material property changes is monitored and obtained in real time.

[0039] Among them, when the five-axis robot processes the footwear according to the established processing path and action sequence, it will use a variety of methods to monitor and obtain possible processing deviations in real time. First, force sensors and temperature sensors are installed on the tool. The force sensor can detect the change in cutting force on the tool during the cutting process. If the tool wears more, the cutting force will usually increase. The temperature sensor is used to monitor the working temperature of the tool. Too high a temperature may mean that the tool is severely worn or the friction during the processing is abnormal. At the same time, a laser measuring instrument or a machine vision system is used to monitor the processing size and shape of the footwear. These devices can obtain the precise size and shape data of the processing part in real time and compare it with the preset standard. If it is found that the actual processing size deviates from the standard size, or the shape does not meet the requirements, it can be judged that a processing deviation has occurred. The monitoring of changes in material properties is achieved by analyzing the sound and vibration signals during the processing process. Different materials will produce different sound and vibration characteristics during processing. When the material properties change, such as a sudden increase in the hardness of the material, the frequency, amplitude and other parameters of the sound and vibration will also change accordingly. In addition, during the processing process, the chips generated by the processing are regularly collected for analysis. The shape, size and color of the chips can reflect the wear of the tool and the stability of the processing process. For example, when engraving the soles of leather shoes, the force sensor detects a sudden increase in cutting force, and the machine vision system finds that the depth of the engraved pattern is shallower than the preset depth. Combined with the analysis of the chips, it is determined that the processing deviation may be caused by tool wear.

[0040] Specifically, in some possible embodiments, the five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the step of real-time monitoring and obtaining processing deviations caused by factors such as tool wear and material property changes includes: Acquiring real-time status data of the footwear when it is processed by monitoring with a variety of sensors configured on the five-axis robot; Comparing the state data with preset standard data or models and feeding back the results to obtain test results, which include analyzing the consistency between the actual material removal amount and the expected amount, the deviation between the processed shape and the actual shoe model, the degree of tool wear and the change in material properties; Based on the test results, re-evaluate the processing of the footwear and obtain an evaluation result; Update the actual shoe model and determine a new processing strategy according to the evaluation result, and dynamically adjust the processing path and the action sequence; The above-mentioned cycle of monitoring, feedback, evaluation and adjustment is continuously carried out in subsequent processing to obtain and deal with processing deviations caused by various factors in real time.

[0041] Among them, the five-axis robot is equipped with a variety of high-precision sensors, such as force sensors, displacement sensors, and visual sensors. During the processing process, the force sensor monitors the cutting force applied by the tool on the shoe material in real time, the displacement sensor obtains the position change of the tool and the shoe, and the visual sensor captures the shape and processing traces of the shoe surface, so as to obtain the real-time status data of the shoe when it is processed. These status data will be quickly transmitted to the data processing center and compared with the preset standard data or model. For example, the actual measured material removal amount is compared with the pre-set expected value to analyze the consistency between the two. At the same time, the deviation between the processed shape and the actual shoe model is detected through image recognition and three-dimensional model comparison technology. In addition, the tool force and wear characteristic data collected by the sensor are used to evaluate the degree of tool wear, and the change of material properties is judged based on the change data of the material's hardness, elasticity and other characteristics. Based on the test results obtained from the above comparison and analysis, the processing of the shoe is re-evaluated. For example, if the actual material removal amount is less than expected, it may mean that the tool is severely worn or the processing parameters are set unreasonably; if the processing shape deviation is large, it may be that there is a problem with the processing path or action sequence. Thus, a comprehensive evaluation result is obtained. According to the evaluation results, the actual footwear model is updated. For example, if some areas are found to be under-processed, they are marked in the model accordingly. At the same time, new processing strategies are determined, such as changing tools, adjusting processing parameters, optimizing processing paths and action sequences, etc., to dynamically adapt to changes in processing. In the subsequent processing process, the above-mentioned monitoring, feedback, evaluation and adjustment cycle is continued. After each processing step is completed, a new round of monitoring and analysis is immediately carried out to obtain and process processing deviations caused by tool wear, changes in material properties and various other factors in real time to ensure that the accuracy and quality of footwear processing always meet the requirements. For example, in the process of processing the sole of a sports shoe, the initial inspection found that the engraving depth of the sole pattern was not enough. After evaluation, it was determined that it was caused by tool wear. The model is then updated, the tool is replaced, and the processing path and action sequence are adjusted. In the subsequent processing, continuous monitoring is carried out. If deviations occur again, the cycle adjustment continues until high-quality processing is completed.

[0042] Specifically, in some possible embodiments, the method includes: The end effector of the five-axis robot and the shoe part being processed are scanned and monitored in real time by the 3D vision sensor to obtain actual processing position information; Comparing and analyzing the actual processing position information with the preset processing position to determine whether the processing position of the five-axis robot is accurate; If the processing position of the five-axis robot is inaccurate, the deviation information is promptly fed back to the control system of the five-axis robot through the 3D vision sensor, and the control system then adjusts the processing path and action sequence of the five-axis robot to correct the deviation; If the processing position of the five-axis robot is accurate, the five-axis robot continues to perform processing operations according to the predetermined processing path and action sequence.

[0043] Among them, at the footwear processing site, the 3D vision sensor is installed at a position where the end effector of the five-axis robot and the footwear parts being processed can be clearly observed. When the five-axis robot starts to process the footwear, the 3D vision sensor starts the real-time scanning function. Through structured light, laser scanning or other technologies, the 3D vision sensor quickly obtains the actual processing position information such as the spatial coordinates and shape of the end effector and the footwear processing part, and transmits these data to the data processing unit. The data processing unit compares and analyzes the received actual processing position information with the preset processing position data. The position deviation between the two is calculated through complex algorithms, including the offset in the X, Y, and Z axis directions in three-dimensional space and the angle deviation. If it is determined that the processing position of the five-axis robot is inaccurate, the 3D vision sensor will immediately feed back detailed deviation information, such as the value and direction of the deviation, to the control system of the five-axis robot. After receiving the deviation information, the control system quickly calculates the new processing path and action sequence according to the preset adjustment strategy and algorithm, and adjusts the motion trajectory and action of the five-axis robot in real time to correct the deviation. For example, if the end effector is found to be in a lower position when stitching the upper, the deviation information will be fed back to the control system, which will then adjust the robot's arm height and stitching action to return it to the correct processing position. If the five-axis robot's processing position is accurate, the control system will not make any adjustments, and the five-axis robot will continue to perform processing operations stably according to the original established processing path and action sequence until the current processing step or the entire processing task is completed.

[0044] Specifically, in some possible embodiments, the scanning frequency and range are pre-set according to the material and shape of the footwear, and the steps of performing a full-scale scanning on the footwear on the production line include: Obtaining and analyzing the material and shape of the footwear; Based on the material and shape, determining the material properties and shape characteristics of the footwear; Determining a scanning frequency of the 3D vision sensor based on reflectivity and transparency of the material properties and complexity of the shape characteristics; Acquiring shape characteristics and preset key part information of the footwear; Determine the scanning range of the 3D vision sensor so that the scanning range can completely cover all parts of the shoe, and at the same time perform a more concentrated and detailed scanning on the key parts corresponding to the key part information; The 3D vision sensor is activated to perform an all-round scan of the shoes continuously moving on the production line. During the scanning process, the quality and integrity of the scanned data are monitored in real time.

[0045] Among them, at the entrance of the production line, material detection equipment and shape measurement devices are set. The material detection equipment obtains the material composition of the footwear through spectral analysis, physical contact and other methods, so as to determine its reflectivity and transparency and other characteristics. The shape measurement device uses image recognition, laser ranging and other technologies to obtain the shape parameters of the footwear such as length, width, height, curve curvature, and conduct detailed analysis. Based on the acquired material and shape information, the system determines the appropriate scanning frequency according to the pre-established correspondence between material characteristics and scanning frequency, as well as the association rules between shape characteristics and scanning frequency. For example, for materials with strong reflectivity and high transparency, such as some bright plastic materials, the scanning frequency is set to a higher value; for footwear with complex shapes and many details, such as boots with complex decorations, the scanning frequency is also increased accordingly. Next, the specific shape characteristics of the footwear are obtained, such as the shape of the toe, the height and shape of the heel, the curve of the upper, etc., and the preset key parts information is referred to, such as the connection between the sole and the upper, the force point of the toe, etc. Then, according to the overall shape and size of the footwear and the distribution of key parts, the scanning range of the 3D vision sensor is determined. Ensure that the scanning range not only covers the entire shoe outline, including conventional parts such as the sole, upper, and upper, but also scans the key parts more intensively and carefully to obtain more accurate data. When the footwear moves continuously on the production line, start the 3D vision sensor to start a full-scale scan. During the scanning process, the system analyzes the received scan data in real time and monitors the quality of the scan data through indicators such as data clarity, integrity, and noise level. If the data is found to be missing or the quality is not up to standard, adjust the sensor parameters or rescan in time to ensure that high-quality and complete scan data is obtained. For example, when scanning a pair of women's high heels, material detection determines that the upper is made of silk and has a certain reflectivity, and shape measurement shows that the heel is slender and complex in shape. Based on this, the system sets a higher scanning frequency and determines that the scanning range should focus on covering the decorative parts of the heel and upper. During the scanning process, real-time monitoring found that the data of the heel part was slightly blurred, and the sensor angle and parameters were immediately adjusted to rescan the heel to ensure the quality and integrity of the data.

[0046] Specifically, in some possible embodiments, the method further includes: After the processing of the footwear is completed, the 3D vision sensor is started to perform an all-round scan of the footwear to detect defects on the surface and inside of the footwear, and obtain defect detection results; Generate defect information according to the defect detection result, wherein the defect information includes the type, location, size and shape of the defect; The defect information is sent to the control system of the five-axis robot through a data transmission channel; after receiving the defect information, the five-axis robot control system plans a repair action and path according to the characteristics and location of the defect; After the repair is completed, the repaired part is scanned again by the 3D vision sensor to evaluate the repair effect.

[0047] If the evaluated repair effect does not meet the requirements, repeat the above steps to perform repair again, and the evaluated repair effect meets the requirements.

[0048] Among them, when the footwear processing is completed, the 3D vision sensor is restarted to perform a full-scale high-precision scan of the footwear. The sensor uses structured light, laser or other advanced scanning technologies to capture the subtle features of the surface and interior of the footwear. Through complex image analysis algorithms and machine learning models, the scanned data is processed and analyzed to detect possible defects in the footwear. These algorithms and models can identify various types of defects such as scratches, wear, color difference, dents, etc., and obtain accurate defect detection results. Based on the detection results, detailed defect information is generated. Clearly record the type of defect, such as scratches or dents; accurately mark the location of the defect, such as the specific coordinates of the toe, heel or upper; accurately measure the size of the defect, such as length, width and depth; and describe the shape of the defect, such as linear, circular or irregular. The generated defect information is quickly sent to the control system of the five-axis robot through a high-speed and stable data transmission channel, such as Ethernet or a dedicated industrial communication protocol. After receiving the defect information, the control system of the five-axis robot uses the pre-set program and algorithm, combined with the robot's kinematic model, to plan precise repair actions and paths based on the characteristics of the defect, such as size, shape and depth, as well as the location of the defect. The repair action may include grinding, filling, painting, etc., and the path ensures that the robot can accurately reach the defect location and perform effective repair operations. After the repair is completed, the 3D vision sensor scans the repaired part again. The repair effect is evaluated by comparing it with the original standard data or the previous flawless model. If the evaluation of the repair effect does not meet the requirements, for example, there is still color difference or uneven surface in the repaired part, the system will automatically repeat the above steps of detection, generation of defect information, sending to the control system, planning of repair actions and paths, and repair again until the evaluation of the repair effect meets the requirements. For example, after processing a pair of leather shoes, the 3D vision sensor detected a tiny scratch on the upper. The generated defect information shows that the scratch type is linear, located in the middle of the upper, about 2 cm long, and shallow in depth. This information is sent to the control system, which plans the sanding and painting repair actions and the corresponding paths. After the repair is completed, another scan and evaluation is performed to find that there is still color difference, so the repair process is repeated until the repair effect fully meets the requirements.

[0049] Specifically, in some possible embodiments, if the evaluation and repair effect does not meet the requirements, the above steps are repeated to perform repair again, and the steps after the evaluation and repair effect meets the requirements include: Collect the shoe data after repair, including the original data with defects, the repair process data and the final data after repair, and mark the defect type, location, severity, etc. in these data in detail; Using the labeled data to train a deep learning model, so that the deep learning can learn the relationship between the characteristics of the defect and the repair method and effect; After the repair is completed, the footwear is scanned again by the 3D vision sensor to obtain detailed information after the repair; Extracting defect-related features from the repaired details; The extracted features are input into the trained deep learning model to predict the cause of the defect and the factors in the processing link that may cause it, and obtain the prediction results; Generate a detailed feedback report based on the prediction results, including possible causes of defects, related processing parameters and actions, etc.; Transmitting the feedback report to the control system of the five-axis robot; After the five-axis robot adjusts the processing parameters and processing actions according to the feedback report, new processing result data is obtained, and the new processing result data is collected again and used to update the deep learning model.

[0050] Among them, when the evaluation of the repair effect meets the requirements, first of all, all kinds of data of the repaired shoes are comprehensively collected. This includes obtaining the original scanning data when the defects exist, the data of the robot's movements and tools used during the entire repair process, and the final state data after the repair is completed. Then, professionals will make detailed and accurate annotations of the defect types (such as scratches, damage, deformation, etc.), locations (accurate to the specific parts of the shoes), and severity (mild, moderate, severe, etc.) in these data. The annotated data is used to train the deep learning model. During the training process, the model learns the complex relationship between various features of the defects (such as shape, color, texture, etc.) and the repair methods (polishing degree, filling materials, paint color, etc.) and the final repair effect through a large amount of data. In subsequent production, every time a pair of shoes is repaired, the 3D vision sensor will scan the shoes again to obtain detailed information after the repair, including surface flatness, color consistency, texture clarity, etc. From these detailed information, feature extraction algorithms are used to extract key features related to the defects, such as color change features of the defective parts, surface roughness features, etc. The extracted features are input into the previously trained deep learning model. Based on the learned knowledge and relationships, the model predicts the cause of the defect, such as excessive tool pressure, excessive processing speed, or quality problems of the material itself during processing, and predicts the specific processing links and related factors that may cause it. Based on the prediction results, a detailed feedback report is generated. The report clearly states the possible causes of the defect, such as "in the sole grinding process, the tool speed is too high, resulting in uneven wear", and lists the relevant processing parameters (such as speed, pressure, temperature, etc.) and actions (such as feed angle, moving speed, etc.). This feedback report is transmitted to the control system of the five-axis robot through a high-speed data transmission channel. After receiving the feedback report, the five-axis robot adjusts the corresponding processing parameters and processing actions according to the suggestions in the report. For example, reduce the tool speed, adjust the pressure value, etc. After the processing is completed, the new processing result data (including various parameters in the processing process and the status of the final product) is collected again. These new data are used to update the deep learning model so that it can continuously learn and improve, improve the accuracy of predicting the cause of the defect and optimize the processing process. For example, for a batch of sports shoes with similar defects, after processing and analysis in the above process, it was found that the cause was uneven stitch density in the upper stitching process. After the feedback report was transmitted to the control system, the five-axis robot adjusted the stitching parameters. The new processing result data was collected and used to update the model, so that the model can more accurately predict and prevent similar defects in the future.

[0051] Specifically, in some possible embodiments, based on the motion accuracy limitation and possible interference of the five-axis robot, the step of adjusting and optimizing the processing path and the action sequence in real time includes: Acquire the current motion accuracy data of the five-axis robot in real time, including position accuracy, posture accuracy, etc., and compare it with the preset accuracy requirements; Monitoring the spatial positional relationship between the components of the five-axis robot, the tool, and the footwear being processed, and predicting whether there is a risk of interference through simulation and calculation; Conduct a detailed analysis of the currently planned machining path to evaluate the movement difficulty and impact on accuracy of each point on the machining path; Check the starting and ending states, speed changes, acceleration changes, etc. of each action in the action sequence, and analyze the smoothness of each action connection; Determine the adjustment strategy based on the accuracy and interference detection results. If the accuracy does not meet the standard or there is a risk of interference, modify the processing path and optimize the action sequence. If the accuracy meets the standard and there is no risk of interference but there is room for optimization, further fine-tune the processing path and the action sequence. regenerating the processing path and the action sequence according to the determined adjustment strategy; The re-planned processing path and the action sequence are quickly simulated and verified, and applied to actual processing in real time after passing the verification.

[0052] Among them, during the process of shoe processing by the five-axis robot, the current motion accuracy data, including precise position coordinates and posture angles, are obtained in real time through sensors such as high-precision encoders and gyroscopes installed on the robot joints and end effectors. These data are quickly transmitted to the control system and compared and analyzed with the pre-set accuracy requirements. At the same time, computer simulation software and real-time calculation modules are used to continuously monitor the spatial position relationship between the components of the five-axis robot (such as robotic arms, joints), tools (such as cutters, fixtures), and the shoes being processed. By establishing a three-dimensional model and kinematic equations, it is predicted whether there is a risk of collision or interference in the next movement. For the currently planned processing path, the path analysis algorithm is used for detailed analysis. Evaluate the movement difficulty of each point on the path, such as sharp turns, steep slopes, etc., and the impact of these points on accuracy, such as whether it is easy to cause position deviation or posture instability. When checking the action sequence, carefully analyze the starting and ending states of each action, including the initial position, final position and posture. Monitor the curve of speed change to ensure that the speed increases and decreases smoothly and avoid sudden stops and starts. Analyze the change of acceleration to prevent excessive acceleration from adversely affecting the robot structure and processing quality. Determine the specific adjustment strategy based on the results of accuracy comparison and interference detection. If it is found that the accuracy does not meet the preset standard or there is a risk of interference, immediately re-plan the processing path, such as changing the direction of the path, adding transition points, etc., and optimize the action sequence, adjust the speed, acceleration and time interval of the action. If the accuracy meets the standard and there is no risk of interference, but there is still room for optimization, such as reducing the movement time or reducing energy consumption, make subtle adjustments to the processing path and action sequence, such as fine-tuning the coordinates of the path points or optimizing the execution order of the actions. According to the determined adjustment strategy, the processing path and action sequence are regenerated using the path planning software and action programming tools. The newly generated path and sequence will be imported into the simulation verification system, and by quickly simulating the actual processing process, check whether there are new problems or potential risks. Once the simulation verification is passed, the new processing path and action sequence will be transmitted to the control system of the five-axis robot in real time and immediately applied to the actual processing to achieve continuous optimization and improvement of the processing process. For example, when processing a pair of high-heeled shoes with a special shape, the real-time position accuracy data shows that the position deviation of a certain processing point exceeds the allowable range, and at the same time, it is detected that there is a risk of interference between the tool and the heel decoration. The system then modifies the processing path to avoid the decoration, optimizes the action sequence, and reduces the processing speed to improve accuracy. After the regenerated path and sequence are verified by simulation, they are applied to actual processing to ensure the processing quality and production efficiency of high-heeled shoes.

[0053] Another embodiment of the present application provides a footwear processing system based on a three-dimensional model, wherein Figure 2 , a footwear processing system based on a three-dimensional model, comprising: Scanning frequency and range preset module 100: obtain the material and shape of the footwear, and preset the scanning frequency and range of the 3D vision sensor based on the material and shape; The three-dimensional data acquisition module 200 performs an all-round scan on the shoes on the production line to obtain the three-dimensional data of each part of the shoes, and monitors and compensates for the data error caused by material deformation in real time during the scanning process; The actual shoe model construction module 300 uses an image processing and modeling algorithm adapted to the characteristics of the shoe data and optimized to pre-process and extract features of the three-dimensional data, compare and verify the three-dimensional data with a standard shoe model, and construct an actual shoe model of the shoe; The actual shoe model transmission module 400 transmits the constructed actual shoe model and related parameters and instructions to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; Processing path and action sequence planning module 500: The five-axis robot control system receives and parses the data instructions, and plans the precise processing path and action sequence according to the information such as the shoe shape, size, material distribution, etc. contained in the actual shoe model and in combination with the pre-optimized processing parameter library; Processing path and action sequence optimization module 600: based on the motion accuracy limit and possible interference of the five-axis robot, real-time adjustment and optimization of the processing path and the action sequence; Processing deviation real-time monitoring module 700: The five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by tool wear, material property changes, etc. is monitored and obtained in real time.

[0054] The present embodiment provides a shoe processing system based on a three-dimensional model. Due to the functions of each module itself and the logical connection between each other, it can implement the various steps of the aforementioned embodiment, and thus can achieve the same technical effect as the aforementioned embodiment. The principle analysis can be found in the relevant description of the steps of the aforementioned shoe processing method based on a three-dimensional model, which will not be repeated here.

[0055] An embodiment of the present application further provides a three-dimensional model-based shoe processing device, including a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the above-mentioned three-dimensional model-based shoe processing method.

[0056] An embodiment of the present application further provides a storage medium, in which a computer program that can be loaded by a processor and execute the above-mentioned shoe processing method based on a three-dimensional model is stored.

[0057] The storage medium provided in this embodiment can achieve the same technical effects as the aforementioned embodiments because the computer program therein implements the various steps of the aforementioned embodiments after being loaded and run on the processor. The principle analysis can be found in the relevant description of the aforementioned method steps, which will not be repeated here.

[0058] The storage medium includes, for example, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0059] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0060] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0061] In addition, the term "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, "plurality" means at least two, such as two, three, etc., unless otherwise clearly and specifically defined, and is only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features.

[0062] Thus, any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0063] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A footwear processing method based on a three-dimensional model, characterized in that: include: Acquire the material and shape of the footwear, and pre-set the scanning frequency and range of the 3D vision sensor based on the material and shape; Performing a full-scale scan of the shoes on the production line to obtain the three-dimensional data of each part of the shoes, and monitoring and compensating for the data errors caused by material deformation in real time during the scanning process; Using an image processing and modeling algorithm that is adapted to the characteristics of the shoe data and optimized, preprocessing and feature extraction are performed on the three-dimensional data, and the three-dimensional data is compared and verified with a standard shoe model to construct an actual shoe model of the shoe; The constructed actual shoe model and related parameters and instructions are transmitted to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; The five-axis robot control system receives and analyzes the data instructions, and plans the precise processing path and action sequence according to the shoe shape, size, and material distribution information contained in the actual shoe model and in combination with the pre-optimized processing parameter library; Based on the motion accuracy limitation and possible interference of the five-axis robot, the processing path and the action sequence are adjusted and optimized in real time; The five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by tool wear and material property change is monitored and obtained in real time.

2. A footwear processing method based on a three-dimensional model according to claim 1, characterized in that: The five-axis robot processes the footwear according to the processing path and action sequence. During the processing, the steps of real-time monitoring and obtaining the processing deviation caused by tool wear and material property change factors include: Acquiring real-time status data of the footwear when it is processed by monitoring with a variety of sensors configured on the five-axis robot; Comparing the state data with preset standard data or models and feeding back the results to obtain test results, which include analyzing the consistency between the actual material removal amount and the expected amount, the deviation between the processed shape and the actual shoe model, the degree of tool wear and the change in material properties; Based on the test results, re-evaluate the processing of the footwear and obtain an evaluation result; Update the actual shoe model and determine a new processing strategy according to the evaluation result, and dynamically adjust the processing path and the action sequence; The above-mentioned cycle of monitoring, feedback, evaluation and adjustment is continuously carried out in subsequent processing to obtain and deal with processing deviations caused by various factors in real time.

3. A footwear processing method based on a three-dimensional model according to claim 1, characterized in that: The method comprises: The end effector of the five-axis robot and the shoe part being processed are scanned and monitored in real time by the 3D vision sensor to obtain actual processing position information; Comparing and analyzing the actual processing position information with the preset processing position to determine whether the processing position of the five-axis robot is accurate; If the processing position of the five-axis robot is inaccurate, the deviation information is promptly fed back to the control system of the five-axis robot through the 3D vision sensor, and the control system then adjusts the processing path and action sequence of the five-axis robot to correct the deviation; If the processing position of the five-axis robot is accurate, the five-axis robot continues to perform processing operations according to the predetermined processing path and action sequence.

4. A footwear processing method based on a three-dimensional model according to claim 1, characterized in that: The scanning frequency and range are pre-set according to the material and shape of the shoes. The steps for performing a full-scale scanning of the shoes on the production line include: Obtaining and analyzing the material and shape of the footwear; Based on the material and shape, determining the material properties and shape characteristics of the footwear; Determining a scanning frequency of the 3D vision sensor based on reflectivity and transparency of the material properties and complexity of the shape characteristics; Acquiring shape characteristics and preset key part information of the footwear; Determine the scanning range of the 3D vision sensor so that the scanning range can completely cover all parts of the shoe, and at the same time perform a more concentrated and detailed scanning on the key parts corresponding to the key part information; The 3D vision sensor is activated to perform an all-round scan of the shoes continuously moving on the production line. During the scanning process, the quality and integrity of the scanned data are monitored in real time.

5. The shoe processing method based on three-dimensional model according to claim 1, characterized in that: The method further comprises: After the processing of the footwear is completed, the 3D vision sensor is started to perform an all-round scan of the footwear to detect defects on the surface and inside of the footwear, and obtain defect detection results; Generate defect information according to the defect detection result, wherein the defect information includes the type, location, size and shape of the defect; The defect information is sent to the control system of the five-axis robot through a data transmission channel; after receiving the defect information, the five-axis robot control system plans a repair action and path according to the characteristics and location of the defect; After the repair is completed, the repaired part is scanned again by the 3D vision sensor to evaluate the repair effect; If the evaluated repair effect does not meet the requirements, repeat the above steps to perform repair again, and the evaluated repair effect meets the requirements.

6. A footwear processing method based on a three-dimensional model according to claim 5, characterized in that: If the evaluation and repair effect does not meet the requirements, repeat the above steps to perform repair again. The steps after the evaluation and repair effect meets the requirements include: Collect the shoe data after repair, including the original data with defects, the repair process data and the final data after repair, and mark the defect type, location and severity in these data in detail; Using the labeled data to train a deep learning model, so that the deep learning can learn the relationship between the characteristics of the defect and the repair method and effect; After the repair is completed, the footwear is scanned again by the 3D vision sensor to obtain detailed information after the repair; Extracting defect-related features from the repaired details; The extracted features are input into the trained deep learning model to predict the cause of the defect and the factors in the processing link that may cause it, and obtain the prediction results; Generate a detailed feedback report based on the prediction results, the feedback including possible causes of defects, related processing parameters and actions; Transmitting the feedback report to the control system of the five-axis robot; After the five-axis robot adjusts the processing parameters and processing actions according to the feedback report, new processing result data is obtained, and the new processing result data is collected again and used to update the deep learning model.

7. A footwear processing method based on a three-dimensional model according to claim 1, characterized in that: Based on the motion accuracy limitation and possible interference of the five-axis robot, the steps of adjusting and optimizing the processing path and the action sequence in real time include: Acquire the current motion accuracy data of the five-axis robot in real time, including position accuracy and posture accuracy, and compare them with the preset accuracy requirements; Monitoring the spatial positional relationship between the components of the five-axis robot, the tools, and the footwear being processed, and predicting whether there is a risk of interference through simulation and calculation; Conduct a detailed analysis of the currently planned machining path to evaluate the movement difficulty and impact on accuracy of each point on the machining path; Check the starting and ending states, speed changes, and acceleration changes of each action in the action sequence, and analyze the smoothness of the connection between each action; Determine the adjustment strategy based on the accuracy and interference detection results. If the accuracy does not meet the standard or there is a risk of interference, modify the processing path and optimize the action sequence. If the accuracy meets the standard and there is no risk of interference but there is room for optimization, further fine-tune the processing path and the action sequence. Regenerate the processing path and the action sequence according to the determined adjustment strategy; The re-planned processing path and the action sequence are quickly simulated and verified, and applied to actual processing in real time after passing the verification.

8. A footwear processing system based on a three-dimensional model, characterized in that: include: Scanning frequency and range preset module: obtains the material and shape of the footwear, and pre-sets the scanning frequency and range of the 3D vision sensor based on the material and shape; 3D data acquisition module: performs all-round scanning of the shoes on the production line to obtain 3D data of each part of the shoes, and monitors and compensates for data errors caused by material deformation in real time during the scanning process; Actual shoe model construction module: using image processing and modeling algorithms adapted to the characteristics of the shoe data and optimized to pre-process and extract features from the three-dimensional data, comparing and verifying the three-dimensional data with a standard shoe model, and constructing an actual shoe model of the shoe; Actual shoe model transmission module: transmits the constructed actual shoe model and related parameters and instructions to the control system of the five-axis robot in real time through a data transmission channel with redundancy and error correction mechanism; Processing path and action sequence planning module: The five-axis robot control system receives and analyzes data instructions, and plans precise processing paths and action sequences according to the shoe shape, size, and material distribution information contained in the actual shoe model and in combination with a pre-optimized processing parameter library; Processing path and action sequence optimization module: based on the motion accuracy limit and possible interference of the five-axis robot, the processing path and the action sequence are adjusted and optimized in real time; Processing deviation real-time monitoring module: the five-axis robot processes the footwear according to the processing path and action sequence, and during the processing, the processing deviation caused by tool wear and material property changes is monitored and obtained in real time.

9. A shoe processing device based on a three-dimensional model, characterized in that: include: A memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executes any one of the three-dimensional model-based shoe processing methods according to claims 1-7.

10. A storage medium, characterized in that: The computer program is stored which can be loaded by a processor and executes any one of the three-dimensional model-based shoe processing methods of claims 1-7.

Images