A clothing sewing management system and method based on RFID technology

Abstract

The invention discloses a clothing sewing management system and method based on RFID technology, and relates to the technical field of clothing sewing; the clothing sewing management system comprises a plurality of function modules which run in sequence, namely a sewing operation monitoring module, a sewing type determination module, a complexity measurement module, a sampling frequency setting module and a feedback correction module; the technical key points are as follows: the management system designs a setting method of a preliminary sampling inspection frequency value in combination with a process type sequence number and complexity, and realizes the accurate matching of the inspection frequency with the process complexity and type; a dynamic sampling inspection frequency adjustment mechanism based on real-time image recognition effectively improves the abnormality detection and processing capability in the production process, and dynamically adjusts the sampling inspection frequency on the basis of the preliminary sampling inspection frequency, which not only improves the accuracy and efficiency of abnormality detection to a certain extent, but also realizes the optimal configuration of inspection resources.

Description

Technical Field

[0001] The present invention relates to the technical field of garment sewing, and in particular to a garment sewing management system and method based on RFID technology. Background Art

[0002] Garment sewing is the process of combining cloth or other materials into complete garments according to specific design patterns and size requirements through manual or machine sewing techniques. This process covers every link from design conception to finished product production and is the core part of garment production. Workers will draw the design drawings of the garments and select suitable fabrics and accessories. Then, according to the design drawings, tailors or equipment will cut the fabrics into pieces of various shapes. These pieces are finely stitched, including hemming, splicing, installation of zippers, buttons and other details, gradually presenting the outline of the garment. During the sewing process, the proficiency of sewing technology and attention to details directly affect the quality and appearance of the garment. In addition, garment sewing also needs to consider the comfort and practicality of wearing to ensure that the garment is not only beautiful but also durable.

[0003] The existing application publication number is CN109359895A, and the document named "An RFID clothing process guidance system and its guidance method, production management system and method" points out that it includes RFID electronic tags, workstations, and network servers, and information can be transmitted between the various components. The RFID electronic tags are attached to the cut pieces, and the information of the cut pieces is written therein. The workstations include RFID readers and display screens. The RFID readers read the information on the RFID electronic tags and transmit the read cut piece information to the network server. Features: The network server includes a database and a process library. The process library determines the process flow chart according to the style type, divides the process of the clothing style, prepares process guidance materials, and stores the process guidance materials in the database. When the process library receives the cut piece information, it transmits the corresponding process guidance materials to the employee's workstation for guiding the employee's operation; it guides the employee's production in real time to avoid sewing errors, better guide the employee to sew, and improve production efficiency; in the traditional management process, it is also necessary to conduct regular sampling inspections on the semi-finished or finished clothing after sewing to ensure the quality and pass rate of the products. At present, there is no management system for the inspection frequency of clothing products on the market.

[0004] In combination with the above documents and existing technologies, there may be multiple processes in the traditional sewing production process. Different types of processes require different equipment setting procedures or manual technical requirements. The more complex the process, the higher the requirements. Similarly, the probability of abnormalities or problems is also higher. When the semi-finished or finished garments produced by the corresponding process are subsequently sampled and inspected, the number of samples taken each time is also different. The inspection frequency setting in traditional production is blind and lacks scientific basis. At the same time, the allocation of quality inspection resources is not reasonable enough, and it is impossible to effectively respond to the high requirements of complex processes for quality control. There may be problems of excessive or insufficient inspection. Summary of the invention

[0005] 1. Technical issues to be resolved In view of the shortcomings of the prior art, the present invention provides a garment sewing management system and method based on RFID technology, designs a method for setting the preliminary sampling inspection frequency value in combination with the process type serial number and complexity, and realizes the precise matching of the inspection frequency with the process complexity and type; a dynamic sampling inspection frequency adjustment mechanism based on real-time image recognition effectively improves the abnormality detection and processing capabilities in the production process, and dynamically adjusts the sampling inspection frequency based on the preliminary sampling inspection frequency, which not only improves the accuracy and efficiency of abnormality detection to a certain extent, but also realizes the optimal allocation of inspection resources, and solves the problems raised in the background technology.

[0006] (II) Technical solution To achieve the above objectives, the present invention is implemented through the following technical solutions: A garment sewing management system based on RFID technology, the system comprising: The sewing operation monitoring module uses an RFID tag to record the material information of the target piece in advance, and uses the RFID tag to record the sewing information during the sewing process of the target piece; wherein the material information at least includes the origin, batch, and quality of the target piece; and the sewing information at least includes the sewing force and sewing speed; The sewing type determination module is based on RFID technology, combined with machine vision monitoring and data analysis and processing to determine the type of the current sewing process; among which, the types of sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; The complexity measurement module runs a simplified processing mechanism according to the sewing process type determined, inputs the obtained results and the pre-collected operation qualification rate into the corresponding complexity calculation model, and outputs the process complexity under the corresponding sewing process type; The sampling frequency setting module combines the type and complexity of the corresponding sewing process to set the preliminary sampling inspection frequency value, which is used to provide inspection frequency guidance for the semi-finished garments obtained from the corresponding type of sewing process; The feedback correction module performs real-time sampling according to the preliminary sampling inspection frequency value, and performs image recognition in real time within the set sampling times M. When there is no abnormality in the recognition result, the established strategy is executed; when there is an abnormality in the recognition result, the feedback correction strategy is executed to correct the preliminary sampling inspection frequency value.

[0007] Furthermore, a management warning unit can be added to the sewing operation monitoring module to monitor sewing information in real time through RFID technology. When the sewing force fluctuation value within the set time window S exceeds the set fluctuation standard value, a first-level warning signal is triggered; when the real-time sewing speed is reduced to 90% of the set speed, a second-level warning signal is triggered. If any warning signal is received, the equipment will be powered off and maintenance will be dispatched; The value range of S is: at least greater than 0; the suture force fluctuation value within the time window S represents: the absolute value of the difference between the maximum value and the minimum value of the suture force within the time window S.

[0008] Further, the process of combining machine vision monitoring and data analysis and processing is as follows: Machine vision system deployment: Machine vision cameras are deployed on the production line of the sewing process to monitor the sewing process in real time; the machine vision system is used to perform image recognition and analysis on the sewing process to identify the type of sewing operation currently being performed, including any one or more of straight sewing and curved sewing; Data collection and integration: Integrate the data collected by RFID technology and machine vision system to form a process data chain to determine the type of sewing process: When the RFID tag shows that the current workstation is any sewing operation type, and the machine vision system determines for the second time that the workstation is performing this type of sewing operation, the current process is determined to be a single sewing operation process, i.e., process one; When the RFID tag shows that the current workstation performs multiple types of sewing operations in sequence, and the machine vision system recognizes the sequence of the multiple types of sewing operations, it is determined that the current process is a multiple serial sewing operation process, that is, process two; When the RFID tag shows that there are multiple workstations performing sewing operations simultaneously, and the machine vision system monitors that the sewing operations are performed simultaneously, it is determined that the current process is a parallel sewing operation process, that is, process three.

[0009] Furthermore, the simplified processing mechanism of operation is as follows: When the sewing type is straight stitching and curved stitching, the formula used to calculate the difficulty coefficient of straight stitching operation is: , where D1 represents the difficulty coefficient of linear suture operation, Indicates the basic difficulty coefficient, Indicates the complexity adjustment factor corresponding to linear suture; The formula used to calculate the difficulty coefficient of the curve stitching operation is: , where D2 represents the difficulty coefficient of curve stitching operation, Indicates the complexity adjustment factor corresponding to curve stitching; The formula used to calculate the parallel coordination coefficient is as follows: ; ; In the formula, C represents the parallel coordination coefficient, bs represents the number of parallel operations, and pr represents the coordination difficulty factor.

[0010] Further, when it is determined that the sewing process type is a single sewing operation process; Then, the operation process of the corresponding complexity calculation model is as follows: ; In the formula, Cp represents the complexity of the process, P represents the qualified rate of the corresponding type of sewing process, and D represents the corresponding operation difficulty coefficient under the corresponding sewing type; When the sewing process type determined to be multiple serial sewing operation processes; Then, the operation process of the corresponding complexity calculation model is as follows: ; In the formula, represents the operation qualification rate of the i-th operation, represents the difficulty coefficient of the i-th operation, and n represents the total number of operations; When the sewing process type is determined to be a parallel sewing operation process; Then, the operation process of the corresponding complexity calculation model is as follows: .

[0011] Furthermore, after dimensionless processing based on the type and complexity of the corresponding sewing process, the preliminary sampling inspection frequency value is set as follows: ; ; Where F represents the frequency of preliminary sampling inspection. Indicates rounding up, Q_nb indicates the value to be rounded, Indicates the growth coefficient, value range: 0< <1, Tx represents the serial number corresponding to the sewing process type.

[0012] Furthermore, the process of performing image recognition actions in real time is as follows: Data preparation: Collect several normal and abnormal images as training data sets, and preprocess the training data sets; abnormal images include broken, uneven, and fractured threads; preprocessing includes image enhancement and standardization; Model selection and training: CNN is selected as the computer vision model architecture, and the CNN model is trained using the preprocessed training data set. During the training process, the model parameters are adjusted through the loss function. Real-time sampling and image recognition: within the set sampling times M, real-time sampling is performed according to the preliminary sampling inspection frequency value, and the sampled image is input into the trained CNN model for recognition. The CNN model outputs a probability value indicating whether the image is abnormal; Abnormal judgment and processing: preset an abnormal judgment threshold; When the probability value is higher than the abnormality judgment threshold, the image is judged to be abnormal; When the probability value is not higher than the abnormality judgment threshold, the image is judged to be normal.

[0013] Furthermore, the established strategy to be executed is: maintaining the original preliminary sampling inspection frequency value to perform real-time sampling actions.

[0014] Furthermore, the feedback correction strategy implemented is as follows: Construct a dynamic adjustment formula based on the abnormal rate: ; Where F_xz represents the dynamic sampling inspection frequency value, Indicates the adjustment coefficient, the value range is 0< <1, Nc represents the number of abnormalities detected in M ​​samplings, and Nc / M represents the abnormality rate.

[0015] A garment sewing management method based on RFID technology comprises the following steps: S1. Using RFID tags to record material information of target pieces in advance, and using RFID tags to record sewing information during sewing of the target pieces; wherein the material information at least includes the origin, batch, and quality of the target pieces; and the sewing information at least includes sewing force and sewing speed; S2. Based on RFID technology, combined with machine vision monitoring and data analysis and processing, the type of the current sewing process is determined; wherein the types of sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; S3. According to the sewing process type determined, the simplified processing mechanism is run, the obtained result and the pre-collected operation qualification rate are input into the corresponding complexity calculation model established, and the process complexity under the corresponding sewing process type is output; S4. In combination with the type and complexity of the corresponding sewing process, a preliminary sampling inspection frequency value is set to provide inspection frequency guidance for semi-finished garments obtained from the corresponding type of sewing process; S5. Perform real-time sampling based on the initial sampling inspection frequency value. Perform image recognition in real time within the set sampling times M. When the recognition result is normal, execute the established strategy. When the recognition result is abnormal, execute the feedback correction strategy to correct the initial sampling inspection frequency value.

[0016] (III) Beneficial effects The present invention provides a garment sewing management system and method based on RFID technology, which has the following beneficial effects: (1) The solution integrates RFID technology, machine vision monitoring and data analysis and processing, which not only ensures the rapid warehousing, outbound and inventory management of raw materials, but also realizes real-time monitoring and early warning of key parameters in the sewing process, effectively reducing equipment failure rate and production abnormalities; the automated judgment of the sewing process reduces manual intervention and improves production efficiency and accuracy; (2) The scheme realizes the accurate quantification of the complexity of the sewing process and the dynamic adjustment of the sampling inspection frequency, which significantly improves the effectiveness and efficiency of production quality control. According to the sewing process type (single, multiple serial, parallel) and its corresponding complexity, a special calculation model is used to comprehensively consider the operation qualification rate, difficulty coefficient and parallel coordination efficiency, and the complexity of the process is reasonably and effectively evaluated. Furthermore, a method for setting the preliminary sampling inspection frequency value is designed in combination with the process type sequence number and complexity, which realizes the accurate matching of the inspection frequency with the process complexity and type. (3) This solution implements a dynamic sampling inspection frequency adjustment mechanism based on real-time image recognition, which effectively improves the anomaly detection and processing capabilities in the production process. On the basis of the preliminary sampling inspection frequency, the real-time sampling image is accurately identified through the convolutional neural network model, and the sampling inspection frequency is dynamically adjusted according to the anomaly rate in the recognition result. When an anomaly is detected, it can quickly feedback and correct the sampling frequency to ensure a timely response to production anomalies. This not only improves the accuracy and efficiency of anomaly detection, but also realizes the optimal allocation of inspection resources and avoids the problem of excessive or insufficient inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a modular schematic diagram of the management system in the present invention; Figure 2 It is a flow chart of the overall steps of the management method in the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1

[0020] See also Figure 1 ,This embodiment provides a clothing sewing management system based on RFID technology; This solution aims to build an efficient and accurate garment sewing management system by introducing RFID (radio frequency identification) technology. The system covers the entire process from sewing garment materials to finished or semi-finished product inspection. Through real-time monitoring and data analysis, the sampling inspection frequency of batch semi-finished products obtained under different sewing process types is adjusted in a targeted manner, which guarantees production efficiency to a certain extent, and can monitor abnormal situations in time, so as to facilitate the subsequent timely adjustment of related equipment, so as to ensure the effective management of the entire sewing operation process; System functional module design The garment sewing management system includes several functional modules that run in sequence, namely, a sewing operation monitoring module, a sewing type determination module, a complexity measurement module, a sampling frequency setting module and a feedback correction module; The following is a description of each functional module in the system: Sewing operation monitoring module: Using RFID tags to record material information of target pieces in advance, and using RFID tags to record sewing information during the sewing process of the target pieces; wherein the material information at least includes the origin, batch, and quality of the target pieces; and the sewing information at least includes sewing force and sewing speed; Among them, for the use of RFID tags on each target piece, the recorded information can be operated as needed; RFID tags are used to record the origin, batch, quality and other information of the target pieces (i.e. the corresponding raw materials) to achieve rapid warehousing, outbound and inventory management of raw materials; RFID technology is used to achieve full traceability of raw materials to ensure that the quality of raw materials is controllable; at the same time, the accuracy and efficiency of raw material management are improved; During the sewing process, RFID tags are used to record key parameters such as sewing force and sewing speed. Usually, the sewing process is performed using automated equipment. It should be noted that the sewing actions performed in this application are all performed using automated equipment, not manual operations. Therefore, a management warning unit can be added to the sewing operation monitoring module to monitor sewing information in real time through RFID technology. When the sewing force fluctuation value within the set time window S exceeds the set fluctuation standard value, a first-level warning signal is triggered; when the real-time sewing speed is reduced to 90% of the set speed, a second-level warning signal is triggered. When the system receives any warning signal, it will cut off the power supply of the equipment and dispatch maintenance. Among them, the value range of S is: at least greater than 0; The suture force fluctuation value within the time window S indicates that: the absolute value of the difference between the maximum value and the minimum value of the suture force within the time window S is the suture force fluctuation value; The first-level warning signal and the second-level warning signal are only used to distinguish the types of equipment abnormalities to facilitate the subsequent maintenance work. At the same time, different methods are used for warnings for the two types of warning signals. For example: the first-level warning signal uses a yellow strobe light to warn, and the second-level warning signal uses a red strobe light to warn.

[0021] Sewing type determination module: Based on RFID technology, combined with machine vision monitoring and data analysis and processing, the type of current sewing process can be determined; The solution installs RFID readers at the production line workstation nodes, equips each target piece with an RFID tag, and combines the machine vision system to monitor the sewing process in real time. Finally, the collected information is processed by the data analysis system to automatically determine the type of the current sewing process. The specific steps of the sewing type determination module are as follows: 1. Data collection and tag binding Application of RFID technology: Install RFID readers at the workstation nodes of the production line (corresponding to different types of sewing processes), and equip each target piece with an RFID tag. When the target piece passes through the workstation node of the production line or is being processed at the workstation node, the RFID tag communicates wirelessly with the RFID reader to read and transmit the information of the RFID tag, such as the current workstation, process status, production batch and other key information; For example, if a target piece is undergoing a linear stitching operation, its RFID tag will record that the current station of the target piece is "linear stitching station" and the process status is "in progress"; 2. Machine vision monitoring Machine vision system deployment: Machine vision cameras are deployed around the production line of the sewing process to monitor the sewing process in real time; Image recognition and analysis: The machine vision system is used to perform image recognition and analysis on the sewing process to identify the type of sewing operation currently being performed, such as straight line sewing, curve sewing, etc.; For example: the machine vision system captures a target piece undergoing a curve stitching operation and transmits the information for analysis; 3. Data analysis and judgment Data collection and integration: Integrate the data collected by RFID technology and machine vision system to form a complete process data chain; Process type determination Single sewing operation process: If the RFID tag shows that the current station is a specific sewing operation (such as a straight-line sewing station), and the machine vision system confirms that the station is performing the sewing operation, then the current process is judged to be a single sewing operation process; For example: The RFID tag shows that the current station is a "straight-line sewing station", and the machine vision system confirms that the station is performing a straight-line sewing operation, then the current process is judged to be a single sewing operation process; Multiple serial sewing operation procedures: If the RFID tag shows that the current station needs to perform multiple sewing operations in sequence (such as straight stitching first and then curve stitching), and the machine vision system can identify the order of these operations, then the current process is judged to be multiple serial sewing operation procedures; for example: the RFID tag shows that the current station needs to perform straight stitching first and then curve stitching, and the machine vision system monitors that these two operations are performed in sequence, then the current process is judged to be multiple serial sewing operation procedures; Parallel sewing operation process: If the RFID tag shows that there are multiple stations currently performing sewing operations at the same time (such as different machines performing straight sewing and curve sewing at the same time), and the machine vision system can monitor that these operations are performed simultaneously, then the current process is judged to be a parallel sewing operation process; for example: the RFID tag shows that there are currently two stations, one performing straight sewing and the other performing curve sewing, and the machine vision system monitors that these two operations are performed simultaneously, then the current process is judged to be a parallel sewing operation process; Therefore, the types of current sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; Each of the above-mentioned sewing processes corresponds to a workstation, which is operated by the automated sewing equipment debugged at the corresponding workstation, and after each sewing process sews the target pieces, a batch of semi-finished garments are obtained; The technical advantages of the solution are as follows: Real-time and accuracy: Through real-time monitoring of RFID technology and machine vision, real-time and accurate judgment of sewing process types can be achieved; Efficiency: Automated judgment reduces manual intervention and waiting time, and improves production efficiency; Scalability: With the continuous expansion and changes of production lines, RFID readers and tags, machine vision cameras, etc. can be easily added and adjusted to meet new production needs; In summary, by combining the comprehensive application of RFID technology, machine vision and data analysis, automatic judgment of sewing process types can be achieved, thereby improving the automation, informatization and intelligence level of the production line.

[0022] By adopting the above technical solutions, real-time and accurate judgment and automatic management of sewing process types are realized, which significantly improves the efficiency and intelligence level of the production line; The solution integrates RFID technology, machine vision monitoring and data analysis and processing, which not only ensures the rapid storage, delivery and inventory management of raw materials, but also realizes real-time monitoring and early warning of key parameters in the sewing process, effectively reducing equipment failure rate and production abnormalities; at the same time, through the automated judgment of the sewing process, it reduces manual intervention and improves production efficiency and accuracy; in addition, the solution also has good scalability and can easily cope with changes and expansion needs of the production line; In summary, this technical solution successfully solved the problems existing in the traditional sewing production process, such as inaccurate process judgment, low production efficiency, and difficult management, and realized the intelligent identification and management of the sewing process.

[0023] Complexity measurement module: According to the sewing process type determined, the simplified processing mechanism is run, and the obtained results and the pre-collected operation qualification rate are input into the corresponding complexity calculation model established, and the process complexity under the corresponding sewing process type is output, which reflects the system's attention to the current type of sewing process; Among them, the simplified processing mechanism process is as follows: When the sewing type is straight stitching and curved stitching, the formula used to calculate the difficulty coefficient of straight stitching operation is: , where D1 represents the difficulty coefficient of linear suture operation, Indicates the basic difficulty coefficient, and the value is usually 1, which represents the most basic sewing operation difficulty. Indicates the complexity adjustment factor corresponding to linear suture, with a value of 0.2. Although linear suture is relatively simple, it still requires certain skills, so a complexity adjustment factor of 0.2 is added; The formula used to calculate the difficulty coefficient of the curve stitching operation is: , where D2 represents the difficulty coefficient of curve stitching operation, Indicates the complexity adjustment factor corresponding to curve stitching, with a value of 0.5. Curve stitching requires higher technical requirements and is more difficult to operate than linear stitching, so the complexity adjustment factor is set to 0.5; The formula used to calculate the parallel coordination coefficient is as follows: ; ; In the formula, C represents the parallel coordination coefficient, bs represents the number of parallel operations, and pr represents the coordination difficulty factor. When pr=0.1, it means that for each additional parallel operation, the coordination difficulty increases by 10%. It should be noted that the parallel coordination coefficient is used to reflect the difficulty of coordination and synchronization between operations in parallel sewing operations; this setting takes into account the impact of the number of parallel operations on the coordination difficulty and quantifies it through the coordination difficulty factor; although the specific value may vary depending on the actual situation, this setting provides a reasonable starting point.

[0024] The sewing process type determined is: single sewing operation process, that is, process one; Then, the operation process of the corresponding complexity calculation model is as follows: ; In the formula, Cp represents the complexity of the process, P represents the operation qualification rate of the corresponding type of sewing process, that is, the probability of successful completion of the operation, and D represents the corresponding operation difficulty coefficient under the corresponding sewing type; Note: For a single sewing operation, its complexity is directly determined by the pass rate and difficulty coefficient of the operation. The higher the pass rate, the greater the difficulty and the higher the complexity of the process. For example, assuming that the pass rate of the straight stitching operation is 95% and the operation difficulty coefficient is 1.2 (determined by the result of the simplified processing mechanism), the process complexity = 95%×1.2 = 1.14; The sewing process type determined is: multiple serial sewing operation processes, that is, process two; Then, the operation process of the corresponding complexity calculation model is as follows: ; In the formula, represents the operation qualification rate of the i-th operation, represents the difficulty coefficient of the i-th operation, and n represents the total number of operations; Note: For multiple serial sewing operations, the complexity is determined by the weighted average of the pass rate and difficulty coefficient of each operation. The weighted average is used here to consider the importance of different operations in the process. For example, if a process contains two sewing operations, the pass rate of the first operation (straight line sewing) is 95% and the difficulty coefficient is 1.2; the pass rate of the second operation (curve sewing) is 90% and the difficulty coefficient is 1.5; then the process complexity = (95%×1.2 + 90%×1.5) / 2 = 1.185; The sewing process type determined is: parallel sewing operation process, that is, process three; Then, the operation process of the corresponding complexity calculation model is as follows: ; In the formula, C represents the parallel coordination coefficient; Note: For parallel sewing operations, in addition to considering the pass rate and difficulty coefficient of each operation, the complexity also needs to consider the collaborative efficiency between parallel operations; the parallel coordination coefficient is used to adjust the increased complexity caused by parallel operations; for example: assuming that a parallel process contains two identical straight-line sewing operations, the pass rate of each operation is 95%, the difficulty coefficient is 1.2, and the parallel coordination coefficient is 1.1; then the process complexity = (95%×1.2 + 95%×1.2) / 2×1.1=1.254.

[0025] Sampling frequency setting module: Combined with the type and complexity of the corresponding sewing process, the preliminary sampling inspection frequency value is set to provide theoretical guidance for the inspection frequency of semi-finished garments obtained from the corresponding type of sewing process; Among them, combined with the type and complexity of the corresponding sewing process, after dimensionless processing, the preliminary sampling inspection frequency value is set as follows: ; ; Where F represents the frequency of preliminary sampling inspection. Indicates rounding up, Q_nb indicates the value to be rounded, Indicates the growth coefficient, which is a fixed value used to adjust the growth rate differences between different process types. The value range is: 0< <1, Tx represents the serial number corresponding to the sewing process type; It should be noted that, considering the different process types, we can adjust the growth rate through an adjustment parameter bT based on the process type number (where bT>1 to ensure exponential growth); for simplicity, we can assume that bT is a function of the process type number Tx, such as bT=1+ ×Tx, where is a fixed growth coefficient; the above formula is ×Tx to adjust the base of the exponential function, thereby realizing the exponential growth demand for inspection frequency of different process types; as the process complexity Cp increases, the inspection frequency F will increase at an exponential rate related to the process type number Tx; The following is an example parameter setting and calculation results, where we assume that =0.2:

[0026] In this example, as the process complexity Cp and the process type sequence number Tx increase, the inspection frequency F increases exponentially, and the above example is only illustrative data to assist understanding.

[0027] When the preliminary sampling inspection frequency value F=2, it means that the number of semi-finished garments selected for each sampling inspection is 2.

[0028] By adopting the above technical solution, the complexity of the sewing process is accurately quantified and the sampling inspection frequency is dynamically adjusted, which significantly improves the effectiveness and efficiency of production quality control. According to the sewing process type (single, multiple serial, parallel) and its corresponding complexity, the solution comprehensively considers the operation qualification rate, difficulty coefficient and parallel coordination efficiency through a special calculation model to scientifically evaluate the complexity of the process. Furthermore, the process type serial number and complexity are combined for dimensionless processing, and a method for setting the preliminary sampling inspection frequency value is innovatively designed, which realizes the precise matching of the inspection frequency with the process complexity and type. This technical solution not only solves the problem of blind inspection frequency setting and lack of scientific basis in traditional production, but also ensures the rational allocation of quality inspection resources and improves inspection efficiency and accuracy. By setting the inspection frequency with exponential growth, it effectively responds to the high requirements of quality control in complex processes, provides the apparel manufacturing industry with intelligent and refined quality management solutions, and effectively promotes the dual improvement of production efficiency and product quality.

[0029] Feedback correction module: Perform real-time sampling actions according to the preliminary sampling inspection frequency value, and perform image recognition actions in real time within the set sampling times M. When the recognition result is normal, execute the established strategy; when the recognition result is abnormal, execute the feedback correction strategy to correct the preliminary sampling inspection frequency value, so as to obtain and perform subsequent sampling actions according to the dynamic sampling inspection frequency value; The process of performing image recognition in real time is as follows: 1. Data preparation Training data set: Collect a large number of normal and abnormal (such as broken, unstraight, and broken) images as training data sets. These data sets can come from historical production records, simulated abnormal scene shooting, etc. Preprocessing: Preprocess the training data, including steps such as image enhancement, resizing, and standardization; Image enhancement can improve the contrast and clarity of the image, which helps to better distinguish between normal and abnormal images; resizing can adjust the input image to a size suitable for the model input; standardization can make the pixel values ​​of the image fall within a specific range to speed up model training and improve the generalization ability of the model; 2. Model selection and training Model architecture: Choose a suitable computer vision model architecture, such as convolutional neural network (CNN). CNN performs well in the field of image processing and can automatically extract image features, which is suitable for anomaly detection tasks; Training process: Use the preprocessed training data set to train the CNN model. During the training process, the model parameters are adjusted through the loss function and optimization algorithm so that the model can accurately identify normal and abnormal images; 3. Real-time sampling and image recognition Real-time sampling: within the set sampling times M, real-time sampling is performed according to the preliminary sampling inspection frequency value; Image recognition: The sampled image is input into the trained CNN model for recognition. The model will output a probability value to indicate whether the image has an abnormality (such as broken threads, uneven lines, or broken lines). 4. Abnormal judgment and processing Set threshold: Set an abnormality judgment threshold according to business needs. If the abnormal probability value output by the model is higher than the threshold, the image is judged to be abnormal; otherwise, there is no abnormality; Abnormal processing: For the abnormal images identified, corresponding processing measures can be taken, such as marking the abnormal location, issuing an alarm, stopping the machine for inspection, etc.; The established strategies for implementation are as follows: Maintain the original preliminary sampling inspection frequency value to perform real-time sampling actions; The feedback correction strategy implemented is as follows: Construct a dynamic adjustment formula based on the abnormal rate: ; Where F_xz represents the dynamic sampling inspection frequency value, Indicates the adjustment coefficient, which is used to control the influence of abnormal times on the sampling inspection frequency value. The value range is 0< <1, Nc represents the number of abnormalities detected in M ​​samplings, and Nc / M represents the abnormality rate; Logic description: When Nc=0, that is, no abnormality is detected, the corrected sampling inspection frequency value is still F; When N>0, that is, an abnormality is detected, the corrected sampling inspection frequency value will increase, and the degree of increase depends on the abnormality rate and the adjustment coefficient; the higher the abnormality rate, the higher the corrected sampling inspection frequency value, and the larger the adjustment coefficient k, the greater the impact of the number of abnormalities on the sampling inspection frequency value, that is, the faster the corrected sampling inspection frequency value increases.

[0030] By adopting the above technical solution, a dynamic sampling inspection frequency adjustment mechanism based on real-time image recognition is realized, which effectively improves the abnormality detection and processing capabilities in the production process; based on the preliminary sampling inspection frequency, the solution accurately recognizes the real-time sampling image through a convolutional neural network model, and dynamically adjusts the sampling inspection frequency according to the abnormality rate in the recognition result; when an abnormality is detected, it can quickly feedback and correct the sampling frequency to ensure timely response to production abnormalities; This technical solution not only improves the accuracy and efficiency of anomaly detection, but also achieves the optimal allocation of inspection resources, avoiding the problem of excessive or insufficient inspection; overall, this solution successfully solves the technical difficulties in traditional sampling inspections where the frequency setting is fixed and cannot adapt to dynamic changes in production.

[0031] Example 2

[0032] See also Figure 2 Based on Example 1, this embodiment also provides a garment sewing management method based on RFID technology, including the following specific steps: S1. Using RFID tags to record material information of target pieces in advance, and using RFID tags to record sewing information during sewing of the target pieces; wherein the material information at least includes the origin, batch, and quality of the target pieces; and the sewing information at least includes sewing force and sewing speed; S2. Based on RFID technology, combined with machine vision monitoring and data analysis and processing, the type of the current sewing process is determined; wherein the types of sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; S3. According to the sewing process type determined, the simplified processing mechanism is run, the obtained result and the pre-collected operation qualification rate are input into the corresponding complexity calculation model established, and the process complexity under the corresponding sewing process type is output; S4. In combination with the type and complexity of the corresponding sewing process, a preliminary sampling inspection frequency value is set to provide inspection frequency guidance for semi-finished garments obtained from the corresponding type of sewing process; S5. Perform real-time sampling based on the initial sampling inspection frequency value. Perform image recognition in real time within the set sampling times M. When the recognition result is normal, execute the established strategy. When the recognition result is abnormal, execute the feedback correction strategy to correct the initial sampling inspection frequency value.

[0033] The above sampling inspection frequencies are for a batch of semi-finished garments obtained under different types of processes. For semi-finished products made by automated equipment, there are rarely abnormalities or errors, so fixed-frequency sampling is required. For the sampling inspection frequency, it means the number of semi-finished products inspected at each fixed time point, for example: 3 pieces each time, 6 pieces each time are both sampling inspection frequencies. If the error rate of the automated equipment is lower, the sampling inspection frequency will be reduced, for example: from the original 6 pieces each time to 3 pieces each time.

[0034] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0035] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0036] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A clothing sewing management system based on RFID technology, characterized in that: The system includes: The sewing operation monitoring module uses an RFID tag to record the material information of the target piece in advance, and uses the RFID tag to record the sewing information during the sewing process of the target piece; wherein the material information at least includes the origin, batch, and quality of the target piece; and the sewing information at least includes the sewing force and sewing speed; The sewing type determination module is based on RFID technology, combined with machine vision monitoring and data analysis and processing to determine the type of the current sewing process; among which, the types of sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; The complexity measurement module runs a simplified processing mechanism according to the sewing process type determined, inputs the obtained results and the pre-collected operation qualification rate into the corresponding complexity calculation model, and outputs the process complexity under the corresponding sewing process type; The sampling frequency setting module combines the type and complexity of the corresponding sewing process to set the preliminary sampling inspection frequency value, which is used to provide inspection frequency guidance for the semi-finished garments obtained from the corresponding type of sewing process; The feedback correction module performs real-time sampling according to the preliminary sampling inspection frequency value, and performs image recognition in real time within the set sampling times M. When there is no abnormality in the recognition result, the established strategy is executed; when there is an abnormality in the recognition result, the feedback correction strategy is executed to correct the preliminary sampling inspection frequency value.

2. The garment sewing management system based on RFID technology according to claim 1, characterized in that: A management warning unit can also be added to the sewing operation monitoring module to monitor sewing information in real time through RFID technology. When the sewing force fluctuation value within the set time window S exceeds the set fluctuation standard value, a first-level warning signal is triggered; When the real-time sewing speed is reduced to 90% of the set speed, the secondary warning signal is triggered; If any warning signal is received, the equipment will be powered off and maintenance will be dispatched; The value range of S is: at least greater than 0; the suture force fluctuation value within the time window S represents: the absolute value of the difference between the maximum value and the minimum value of the suture force within the time window S.

3. The garment sewing management system based on RFID technology according to claim 1, characterized in that: The process of combining machine vision monitoring and data analysis and processing is as follows: Machine vision system deployment: Machine vision cameras are deployed on the production line of the sewing process to monitor the sewing process in real time; the machine vision system is used to perform image recognition and analysis on the sewing process to identify the type of sewing operation currently being performed, including any one or more of straight sewing and curved sewing; Data collection and integration: Integrate the data collected by RFID technology and machine vision system to form a process data chain to determine the type of sewing process: When the RFID tag shows that the current workstation is any sewing operation type, and the machine vision system determines for the second time that the workstation is performing this type of sewing operation, the current process is determined to be a single sewing operation process, i.e., process one; When the RFID tag shows that the current workstation performs multiple types of sewing operations in sequence, and the machine vision system recognizes the sequence of the multiple types of sewing operations, it is determined that the current process is a multiple serial sewing operation process, that is, process two; When the RFID tag shows that there are multiple workstations performing sewing operations simultaneously, and the machine vision system monitors that the sewing operations are performed simultaneously, it is determined that the current process is a parallel sewing operation process, that is, process three.

4. The garment sewing management system based on RFID technology according to claim 1, characterized in that: The simplified processing mechanism process is as follows: When the sewing type is straight stitching and curved stitching, the formula used to calculate the difficulty coefficient of straight stitching operation is: , where D1 represents the difficulty coefficient of linear suture operation, Indicates the basic difficulty coefficient, Indicates the complexity adjustment factor corresponding to linear suture; The formula used to calculate the difficulty coefficient of the curve stitching operation is: , where D2 represents the difficulty coefficient of curve stitching operation, Indicates the complexity adjustment factor corresponding to curve stitching; The formula used to calculate the parallel coordination coefficient is as follows: ； ； In the formula, C represents the parallel coordination coefficient, bs represents the number of parallel operations, and pr represents the coordination difficulty factor.

5. The garment sewing management system based on RFID technology according to claim 4 is characterized in that: When the sewing process type determined to be a single sewing operation process; Then, the operation process of the corresponding complexity calculation model is as follows: ； In the formula, Cp represents the complexity of the process, P represents the qualified rate of the corresponding type of sewing process, and D represents the corresponding operation difficulty coefficient under the corresponding sewing type; When the sewing process type determined to be multiple serial sewing operation processes; Then, the operation process of the corresponding complexity calculation model is as follows: ； In the formula, represents the operation qualification rate of the ith operation, represents the difficulty coefficient of the i-th operation, and n represents the total number of operations; When the sewing process type is determined to be a parallel sewing operation process; Then, the operation process of the corresponding complexity calculation model is as follows: 。 6. The garment sewing management system based on RFID technology according to claim 5, characterized in that: Combined with the type and complexity of the corresponding sewing process, after dimensionless processing, the preliminary sampling inspection frequency value is set as follows: ； ； Where F represents the frequency of preliminary sampling inspection. Indicates rounding up, Q_nb indicates the value to be rounded, Indicates the growth coefficient, value range: 0< <1, Tx represents the serial number corresponding to the sewing process type.

7. The garment sewing management system based on RFID technology according to claim 1, characterized in that: The process of performing image recognition actions in real time is as follows: Data preparation: Collect several normal and abnormal images as training data sets, and preprocess the training data sets; abnormal images include broken, uneven, and fractured threads; preprocessing includes image enhancement and standardization; Model selection and training: CNN is selected as the computer vision model architecture, and the CNN model is trained using the preprocessed training data set. During the training process, the model parameters are adjusted through the loss function. Real-time sampling and image recognition: within the set sampling times M, real-time sampling is performed according to the preliminary sampling inspection frequency value, and the sampled image is input into the trained CNN model for recognition. The CNN model outputs a probability value indicating whether the image is abnormal; Abnormal judgment and processing: preset an abnormal judgment threshold; When the probability value is higher than the abnormality judgment threshold, the image is judged to be abnormal; When the probability value is not higher than the abnormality judgment threshold, the image is judged to be normal.

8. The garment sewing management system based on RFID technology according to claim 1, characterized in that: The established strategy to be implemented is: maintaining the original preliminary sampling inspection frequency value to perform real-time sampling actions.

9. The garment sewing management system based on RFID technology according to claim 1, characterized in that: The feedback correction strategy implemented is as follows: Construct a dynamic adjustment formula based on the abnormal rate: ； Where F_xz represents the dynamic sampling inspection frequency value, Indicates the adjustment coefficient, the value range is 0< <1, Nc represents the number of abnormalities detected in M ​​samplings, and Nc / M represents the abnormality rate.

10. A garment sewing management method based on RFID technology, using any system described in claims 1 to 9, characterized in that: The steps include: S1. Using RFID tags to record material information of target pieces in advance, and using RFID tags to record sewing information during sewing of the target pieces; wherein the material information at least includes the origin, batch, and quality of the target pieces; and the sewing information at least includes sewing force and sewing speed; S2. Based on RFID technology, combined with machine vision monitoring and data analysis and processing, the type of the current sewing process is determined; wherein the types of sewing processes include: Process 1: Single sewing operation process; Process 2: Multiple serial sewing operations; Process three: parallel sewing operation process; S3. According to the sewing process type determined, the simplified processing mechanism is run, the obtained result and the pre-collected operation qualification rate are input into the corresponding complexity calculation model established, and the process complexity under the corresponding sewing process type is output; S4. In combination with the type and complexity of the corresponding sewing process, a preliminary sampling inspection frequency value is set to provide inspection frequency guidance for semi-finished garments obtained from the corresponding type of sewing process; S5. Perform real-time sampling based on the initial sampling inspection frequency value. Perform image recognition in real time within the set sampling times M. When the recognition result is normal, execute the established strategy. When the recognition result is abnormal, execute the feedback correction strategy to correct the initial sampling inspection frequency value.

Images