Intelligent laser cutting control method, system and equipment in clothing industry

Abstract

The invention discloses an intelligent laser cutting control method for the clothing industry. The method comprises the following steps: analyzing fabric information; performing fabric identification; actual cutting speed data and seam width data are collected, and whether cutting succeeds or not is judged; executing data acquisition, and transmitting cut data to a server side in real time; executing data cleaning; performing optimization on the data set; executing data preprocessing on the obtained rate optimization data set; laser cutting bed rate regression model training is executed; executing acquisition optimization based on the data of the correlation analysis, and optimizing data acquisition content based on a correlation analysis result; executing edge end station cutting picture data acquisition, and transmitting the data to a server end in real time; laser cutting bed parameter control model deployment is executed; a regression model running on a cloud platform or an edge end carries out regression on the optimal cutting rate according to the currently collected real-time fabric data; and cutting based on the regression rate.

Description

Technical Field

[0001] The present invention relates to the field of computers, and particularly to an intelligent laser cutting control method, system and device for the clothing industry. Background Art

[0002] Laser processing technology is a non-contact processing method, which has the advantages of high precision, high speed, small heat affected zone, etc., and is widely used in the cutting, engraving and drilling of various materials. In the clothing manufacturing industry, by using laser cutting technology, various shapes of fabrics can be quickly and accurately cut according to preset patterns, improving production efficiency.

[0003] Machine vision technology is an interdisciplinary subject involving image processing, computer vision, pattern recognition and other technologies. It uses devices such as cameras to obtain image information of objects, and through computer processing and analysis, extracts useful information to achieve functions such as measurement, recognition and positioning of objects. In laser cutting technology, machine vision technology can be used to perceive the quality of current cutting. Artificial intelligence learning technology is an important branch in the field of artificial intelligence. It enables computer systems to learn from data and obtain intelligence by simulating the learning process of humans. In laser cutting technology, artificial intelligence learning technology can be used to optimize cutting parameters and improve cutting effects.

[0004] Existing laser cutting technologies mainly control the movement of the laser head through preset cutting paths to cut fabrics into predetermined shapes. To ensure cutting quality, manual inspection of the cutting effect is usually required, and if problems are found, cutting parameters need to be manually adjusted. In current production practices, there are certain problems with the use of laser cutting machines by production line workers: the control of cutting speed depends on the production experience of employees, and employees only care about production efficiency and do not care about product life and equipment wear. First, the characteristics of the fabric such as yarn count, gram weight, material, thickness, humidity, brand, etc. will affect the cutting speed. Second, the section leader of cutting usually tells the cutting employees what the maximum power setting of the laser cutting machine is, and employees usually select the maximum cutting speed according to the maximum power to maximize the cutting efficiency of this station. The method for employees to determine the maximum speed is to repeatedly try based on experience, and the fabric loss caused during the trial process is not considered by employees. Summary of the Invention

[0005] One object of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which provides intelligent laser cutting for the clothing industry, especially for customized clothing production, based on precise monitoring and self-perception learning.

[0006] One object of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can timely detect and correct problems in the laser cutting process to improve cutting quality.

[0007] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can automatically adjust cutting parameters according to the characteristics of fabrics, such as material, yarn count, gram weight and thickness, etc., so as to improve the cutting effect.

[0008] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can optimize cutting parameters through artificial intelligence learning technology, so as to improve cutting efficiency and reduce resource waste.

[0009] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can precisely control the cutting effect. Through intelligent monitoring, problems in the cutting process can be discovered and corrected in a timely manner, improving the accuracy and stability of cutting.

[0010] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can provide an adaptive adjustment mechanism. According to factors such as the material, yarn count, gram weight and thickness of the fabric, it performs self-perception learning and scores the effect after each cutting. Through continuous evaluation and comparison, the best matching value is selected to determine the most ideal cutting effect, enabling the laser cutting technology to adapt to fabrics of various different materials and characteristics, and improving the flexibility and adaptability of cutting.

[0011] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can save resources. It automatically adjusts the spot size, height of the laser head and the running speed of the cutting bed head according to the best matching value, ensuring the accuracy and efficiency of the cutting process, while minimizing fabric waste and defects to achieve resource conservation and utilization.

[0012] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can improve quality and efficiency. The cut fabric pieces have advantages such as neat edges and no wire drawing, enhancing the overall aesthetic and quality of the clothing.

[0013] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which can reduce costs. By improving the accuracy and efficiency of the cutting process, reducing the influence of human factors, and lowering labor costs and time costs, the production cost can be reduced, and the competitiveness of the enterprise can be improved.

[0014] One of the objectives of the present invention is to provide an intelligent laser cutting control method, system and device for the clothing industry, which provides an optimization process for the cutting rate. For the data set formed by data collection in this process, learning is carried out on this data set, and the optimal cutting rate for different fabrics can be regressed.

[0015] To achieve at least one inventive object of the present invention, the present invention provides an intelligent laser cutting control method for the clothing industry. The intelligent laser cutting control method for the clothing industry includes the following steps:

[0016] Analyze fabric information;

[0017] Perform fabric recognition to determine whether the feeding is correct. When the feeding is correct, read the fabric information and the laser power of the laser cutting bed;

[0018] Collect actual cutting speed data and cutting width data, and determine whether the cutting is successful;

[0019] Perform data collection, and the cutting data is transmitted to the server side in real time;

[0020] Perform data cleaning;

[0021] Optimize the data set, filter out the data with the largest weight assignment, obtain a weighted data set, and perform processing on the weighted data set: group the data according to fabric material, density, type, and gram weight, and obtain the maximum rate of each group of data; check the number of duplicate data entries. If it exceeds the preset threshold, it is determined that the data is not noise, and the rate of all data in this group is modified to the current rate value to obtain an optimized rate optimization data set;

[0022] Perform data preprocessing on the obtained rate optimization data set;

[0023] Perform training of the laser cutting bed rate regression model;

[0024] Perform acquisition optimization based on correlation analysis data, and optimize the data acquisition content based on the correlation analysis results;

[0025] Perform data collection of cutting pictures at the edge-side workstations, and the data is transmitted to the server side in real time;

[0026] Deploy the laser cutting bed parameter control model;

[0027] Run the regression model on the cloud platform or the edge side, and perform regression on the optimal cutting rate based on the currently collected real-time fabric data; and

[0028] Perform cutting based on the regressed rate.

[0029] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the steps of: controlling the selection of industrial lenses and lighting, and controlling the installation and debugging of the industrial lens and lighting auxiliary modules.

[0030] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the step of: performing the training method of the laser cutting bed rate regression model using XGBoost.

[0031] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the step of: performing data cleaning selected from performing missing value processing, performing duplicate data processing, performing error data correction, performing data standardization, performing data normalization, performing format conversion, performing outlier processing, performing data verification, performing skewness processing.

[0032] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the step of: performing data preprocessing on the rate optimization data set, adding a success data field to all data, setting the value of the success data item to 1 for the data with the maximum weight assignment, and setting the value of the success data item to 0 for the data with a weight assignment that is not the maximum.

[0033] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the step of: observing the cutting effect during the cutting process, determining whether the cutting is completed. If the cutting is completed, perform data collection and construct the laser cutting bed parameter control model. If the cutting is not completed, retrieve the data of the recorded rate debugging process and train the laser cutting bed parameter control model based on the new data set for iteration.

[0034] In some embodiments, the intelligent laser cutting control method for the clothing industry further includes the step of: performing data preprocessing selected from performing data integration, performing data transformation, performing feature engineering, performing data binning, performing class balance processing, performing data skewness processing.

[0035] According to another aspect of the present invention, there is also provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it executes the steps of the intelligent laser cutting control method for the clothing industry.

[0036] According to another aspect of the present invention, there is also provided an intelligent laser cutting control device for the clothing industry, including:

[0037] A memory for storing software application programs,

[0038] A processor for executing the software application programs, and each program of the software application programs correspondingly executes each step of any one of the intelligent laser cutting control methods for the clothing industry.

[0039] According to another aspect of the present invention, there is also provided an intelligent laser cutting control system for the garment industry, which applies the intelligent laser cutting control method for the garment industry. The intelligent laser cutting control system for the garment industry includes a fabric analysis unit, a fabric information identification and reading unit, a cutting judgment unit, a data collection unit, a data cleaning unit, a data set optimization unit, a data preprocessing unit, a laser cutting bed rate regression model training unit, a data collection optimization unit, a lens lighting control unit, an edge-side station cutting picture data collection unit, a laser cutting bed rate regression model deployment unit, a regression unit, and a cutting control unit. Among them, the fabric analysis unit is used to analyze fabric information, the fabric information identification and reading unit is used to perform fabric identification and judge whether the feeding is correct. When the feeding is correct, the fabric information is read. The cutting judgment unit is used to judge whether the cutting is successful. The data collection unit is used to perform data collection, and the cutting data is transmitted to the server end in real time. The data cleaning unit is used to perform data cleaning. The data set optimization unit optimizes the data set. The data preprocessing unit is used to perform preprocessing on the data set optimized by the data set optimization unit. The laser cutting bed rate regression model training unit is used to perform training of the laser cutting bed rate regression model. The data collection optimization unit is used to perform collection optimization based on correlation analysis data, and optimize the data collection content based on the correlation analysis result. The lens lighting control unit is used to control the selection of industrial lenses and lighting, and control the installation and debugging of industrial lenses and lighting auxiliary modules. The edge-side station cutting picture data collection unit is used to perform collection of edge-side station cutting picture data, and the data is transmitted to the server in real time. The laser cutting bed rate regression model deployment unit is used to perform deployment of the laser cutting bed parameter control model. The regression unit is used to run the regression model on the cloud platform or the edge side, and perform regression on the optimal cutting rate according to the currently collected real-time fabric data. The cutting control unit is used to perform cutting based on the regressed rate.

[0040] In some embodiments, the cutting judgment unit includes a cutting judgment module and an indicator variable module. The cutting judgment module obtains the actually collected cutting speed data and the tailor width data, and judges whether the cutting is successful. The indicator variable module outputs a cutting success indicator variable. When the cutting is successful, it outputs a cutting success indicator variable = 1, indicating that the cutting is successful. When the cutting is not successful, it outputs a cutting success indicator variable = 0, indicating that the cutting is not successful.

[0041] In some embodiments, the intelligent laser cutting control system for the garment industry further includes a feature engineering unit for performing feature engineering.

[0042] In some embodiments, the feature engineering unit includes a heat map analysis module, and the heat map analysis module uses a heat map to analyze the correlation between fabric material, fabric density, fabric type, gram weight, cutting bed power, cutting speed, cutting width, and the cutting success indication variable.

[0043] In some embodiments, the feature engineering unit includes a monotonic similarity analysis module, and the monotonic similarity analysis module analyzes the monotonic similarity between fabric parameters and the cutting success indication variable through the Kendall coefficient.

[0044] In some embodiments, the feature engineering unit includes an information gain analysis module and a feature set determination module. The information gain analysis module analyzes the predictability between fabric parameters and the cutting success indication variable through information gain, and the feature set determination module determines the feature set used for predicting the cutting success indication variable based on the predictability analysis of the information gain analysis module. Description of the Drawings

[0045] Figure 1 It is a step flowchart of the laser cutting bed parameter control of an intelligent laser cutting control method for the clothing industry according to an embodiment of the present invention, which illustrates data collection and data weight assignment.

[0046] Figure 2 It is a step flowchart of the laser cutting bed parameter control of the intelligent laser cutting control method for the clothing industry according to the above embodiment of the present invention, which illustrates the optimization of the data set.

[0047] Figure 3 It is a step flowchart of the laser cutting bed parameter control of the intelligent laser cutting control method for the clothing industry according to the above embodiment of the present invention, which describes the marking of data.

[0048] Figure 4 It is a step flowchart of the laser cutting bed parameter control of the intelligent laser cutting control method for the clothing industry according to the above embodiment of the present invention, which illustrates the optimization of the laser cutting bed parameter control model. Detailed Embodiments

[0049] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments in the following description are only examples, and those skilled in the art can think of other obvious variations. The basic principles defined in the following description can be applied to other embodiments, variations, improvements, equivalent solutions, and other technical solutions without departing from the spirit and scope of the present invention.

[0050] It is understood that the term "a" should be construed as "at least one" or "one or more". That is, in one embodiment, the number of an element can be one, while in other embodiments, the number of the element can be multiple. The term "a" should not be construed as a limitation on the quantity.

[0051] The present invention is an invention related to computer programs. The present invention provides a flowchart of an intelligent laser cutting control method for the clothing industry, and elaborates a solution for controlling or processing external or internal objects of a computer by executing a computer program compiled according to the above process based on the computer program processing process to solve the problems proposed by the present invention. Through the intelligent laser cutting control method for the clothing industry of the present invention, it is possible to use a computer system to provide intelligent laser cutting for the clothing industry, especially for customized clothing production, based on precise monitoring and self-perception learning; it is possible to timely discover and correct problems in the laser cutting process to improve the cutting quality; it is possible to automatically adjust cutting parameters according to the characteristics of the fabric, such as material, yarn count, gram weight, and thickness, etc., to improve the cutting effect; it is possible to optimize cutting parameters through artificial intelligence learning technology to improve cutting efficiency and reduce resource waste; it is possible to precisely control the cutting effect, and through intelligent monitoring, timely discover and correct problems in the cutting process to improve the accuracy and stability of cutting; it is possible to provide an adaptive adjustment mechanism, perform self-perception learning according to factors such as the material, yarn count, gram weight, and thickness of the fabric, and score the effect after each cutting. Through continuous evaluation and comparison, the best matching value is selected to determine the most ideal cutting effect, so that the laser cutting technology can adapt to fabrics of various different materials and characteristics, improving the flexibility and adaptability of cutting; it is possible to save resources, automatically adjust the spot size, height of the laser head, and the running speed of the cutting bed head according to the best matching value, ensuring the accuracy and efficiency of the cutting process, while minimizing fabric waste and defects to achieve resource conservation and utilization; it is possible to improve quality and efficiency. The cut fabric pieces have advantages such as neat edges and no wire drawing, enhancing the overall aesthetics and quality of the clothing; it is possible to reduce costs, by improving the accuracy and efficiency of the cutting process, reducing the influence of human factors, reducing labor costs and time costs, thereby reducing production costs and enhancing the competitiveness of enterprises.

[0052] The intelligent laser cutting control method for the clothing industry includes the following laser cutting bed parameter control steps:

[0053] S10: Analyze the fabric information. Preferably, the fabric information comes from the cutting management system;

[0054] S21: Perform fabric identification, determine whether the feeding is correct. When the feeding is correct, read the fabric information and the laser power p of the laser cutting bed i , where the fabric information includes the fabric material mi , fabric density d i , fabric type (yarn count, knitting, weaving, etc.), gram weight g i , hardness h i , flexibility f i , elasticity e i etc.;

[0055] S22: Collect the actual cutting speed v i data, and the cutting width q i data, and determine whether the cutting is successful, where the cutting success indicator variable is c i , that is, c i = 1 indicates successful cutting, c i = 0 indicates unsuccessful cutting;

[0056] S31: Perform data collection, and the cutting data is transmitted to the server side in real time. Preferably, it is transmitted in real time based on the industrial communication protocol;

[0057] S32: Perform data cleaning, where data cleaning includes but is not limited to missing value processing, duplicate data processing, error data correction, data standardization / normalization, format conversion, outlier processing, data verification, skewness processing, etc.;

[0058] S41: Optimize the data set, filter all data with weight = 5 to obtain the data set S c , perform processing on the data set S c : Group the data according to the fabric material m i , fabric density d i , gram weight g i , hardness h i , flexibility f i , elasticity e i and fabric type, that is, those with the same parameters are grouped into one group, and the maximum rate top10 of each group of data is obtained; Starting from top1, check the number of duplicate data. If it exceeds the preset threshold, it is considered that the data is not noise, and the rate of all data in this group is modified to the current rate value (the maximum rate of the data exceeding the threshold) to obtain the new data set S n ;

[0059] S42: Perform data preprocessing on the data set S n : Add a success data field to all data, set the value of the success data item to 1 for all data with weight = 5, and set the value of the success data item to 0 for data with weight not equal to 5;

[0060] S50: Perform feature engineering;

[0061] S60: Perform the training of the laser cutting machine speed regression model, where the XGBoost method is used for training; a speed prediction model is trained, and this model can regress the optimal cutting speed based on the current information faced.

[0062] S70: Perform acquisition optimization based on the data of correlation analysis, and optimize the data acquisition content based on the correlation analysis results, that is, the parameter set for sampling is reduced to CH i The parameters included;

[0063] S80: Control the selection of industrial lenses and lighting, and control the installation and debugging of industrial lens and lighting auxiliary modules;

[0064] S90: Perform the acquisition of cutting picture data at the edge - end workstation, and the data is transmitted to the server in real - time. Preferably, it is transmitted based on the industrial communication protocol;

[0065] S100: Perform the deployment of the laser cutting machine parameter control model, including cloud deployment and edge - end deployment, and the deployment content is the model trained in step S60;

[0066] S110: The regression model running on the cloud platform or the edge - end regresses the optimal cutting speed based on the current real - time fabric data collected;

[0067] S120: Perform cutting based on the regressed speed; during the cutting process, observe the cutting effect. If it can be cut off, perform data acquisition according to the process in step S31, and perform model construction for the subsequent process; if it cannot be cut off, record the data of the speed debugging process of the cutting workstation employees according to the process in S31, and train a new model for iteration after obtaining a new data set.

[0068] Preferably, in step S21, fabric identification is performed, which is specifically implemented as: feeding the fabric based on a barcode scanner, scanning the fabric QR code, and verifying the fabric identification code.

[0069] Preferably, in the preferred embodiment, step S22 is specifically implemented as: installing an industrial lens in the laser cutting module, mainly collecting the actual cutting speed v i and the sewing width q i , the cutting success indication variable c i , that is, c i = 1 indicates successful cutting, and c i = 0 indicates unsuccessful cutting.

[0070] Preferably, step S31 includes the following judgments on the scenario, and performs data acquisition and data weight assignment steps according to the scenario judgment results:

[0071] When it is determined that the scenario is a failed cutting scenario during fabric replacement, execute: Data empowerment in this scenario, adding a weight field weight4;

[0072] When it is determined that the scenario is the scenario of starting cutting after several debugging sessions are completed after fabric replacement, execute: Data empowerment 5 in this scenario;

[0073] When it is determined that the scenario is successful cutting during fabric replacement, but the change in the tailor width exceeds 20% or more, execute: Data empowerment 4 in this scenario;

[0074] When it is determined that the scenario is successful cutting during fabric replacement, but the change in the tailor width is within 20%, execute: Data empowerment 5 for the data in this scenario;

[0075] When it is determined that the scenario is a scenario without fabric replacement, execute: Data empowerment 1;

[0076] When it is determined that the scenario is a scenario of unsuccessful cutting during the debugging process after fabric replacement, execute: Data empowerment 4 for the data in this scenario.

[0077] Among them, the data preprocessing in step S42 includes but is not limited to data integration, data transformation, feature engineering, data binning, class balance processing, data skewness processing, etc.

[0078] Specifically, step S50 further includes the following steps:

[0079] S51: Use a heatmap to analyze the correlation between the aforementioned fabric material, fabric density, fabric type, gram weight, cutting bed power, cutting speed, tailor width q i , and the cutting success indicator variable c i ;

[0080] S52: Analyze the monotonic similarity between these parameters and the cutting success indicator variable c i through the Kendall coefficient;

[0081] S53: Analyze the predictability between these parameters and the cutting success indicator variable c i through information gain;

[0082] S54: Based on the aforementioned predictability analysis, determine the feature set CH i used for predicting the cutting success indicator variable c i .

[0083] Among them, step S10 further includes the steps: constructing a cutting parameter model, which processes the input and updated production practice experience and outputs the cutting parameters of the required cutting data for the current station at the edge side. The cutting parameters include but are not limited to fabric thickness, cutting power, tailor width, main fabric material, secondary fabric material, and fabric gram weight.

[0084] Specifically, in other preferred embodiments of the present invention, the intelligent laser cutting control method for the clothing industry includes the following steps for optimizing the laser cutting machine parameter control model:

[0085] Perform data acquisition;

[0086] Load the laser cutting machine parameter control model;

[0087] Run the laser cutting machine parameter control model;

[0088] Perform speed optimization;

[0089] Set the optimized speed;

[0090] Observe the cutting effect;

[0091] Judge whether the cutting is completed;

[0092] If the judgment result is that the cutting is not completed, perform debugging. After debugging, perform data acquisition to form a new dataset after debugging;

[0093] If the judgment result is that the cutting is completed, perform data acquisition, construct a new dataset, and perform model optimization based on the new dataset.

[0094] Those skilled in the art can understand that the embodiments of the present invention can be provided in the form of a method, a system, or a computer program product. Therefore, the present invention can take the form of an all-hardware embodiment, an all-software embodiment, or an embodiment combining software and hardware.

[0095] The present invention can be embedded in a computer program product, which includes all features enabling the methods described herein to be implemented. The computer program product is contained in one or more computer-readable storage media, which have computer-readable program code contained therein. According to another aspect of the present invention, a computer-readable storage medium is also provided, on which a computer program is stored. When the computer program is executed by a processor, it can execute the steps of the intelligent laser cutting control method for the clothing industry of the present invention. A computer storage medium is a medium in a computer memory for storing a certain discontinuous physical quantity. Computer storage media include, but are not limited to, semiconductors, disk memories, magnetic cores, magnetic drums, magnetic tapes, laser discs, etc. Those skilled in the art can understand that computer storage media are not limited to the foregoing examples, and the foregoing examples are merely examples and not limited to the present invention.

[0096] According to another aspect of the present invention, there is also provided an intelligent laser cutting control device for the clothing industry, which includes: a software application program, a memory for storing the software application program, and a processor for executing the software application program. Each program of the software application program can correspondingly execute the steps in the intelligent laser cutting control method for the clothing industry of the present invention.

[0097] A typical combination of hardware and software can be a general computer system with a computer program. When the program is loaded and executed, it controls the computer system, so that the intelligent laser cutting control method disclosed in the present invention can be executed.

[0098] Those skilled in the art can understand that the intelligent laser cutting control device for the clothing industry can be embodied as a desktop computer, a notebook, a mobile intelligent device, etc. However, the foregoing are merely examples, and it also includes other intelligent devices equipped with the software application program of the present invention.

[0099] Those skilled in the art can understand that the intelligent laser cutting control method for the clothing industry of the present invention can be implemented through hardware, software, or a combination of software and hardware. The present invention can be implemented in a centralized manner in at least one computer system, or be implemented in a decentralized manner by different parts distributed in several interconnected computer systems. Any computer system or other device that can implement the method is applicable. A common combination of software and hardware can be a general computer system installed with a computer program. By installing and executing the program, the computer system is controlled to operate according to the intelligent laser cutting control method for the clothing industry.

[0100] Corresponding to the embodiment of the method of the present invention, according to another aspect of the present invention, there is also provided an intelligent laser cutting control system for the clothing industry, and the intelligent laser cutting control system for the clothing industry is an application of the intelligent laser cutting control method for the clothing industry of the present invention in computer program improvement.

[0101] Through the intelligent laser cutting control system for the clothing industry of the present invention, it is possible to provide intelligent laser cutting for the clothing industry, especially for customized clothing production, based on precise monitoring and self-perception learning; it can timely detect and correct problems in the laser cutting process to improve the cutting quality; it can automatically adjust cutting parameters according to the characteristics of the fabric, such as material, yarn count, gram weight, and thickness, to improve the cutting effect; it can optimize cutting parameters through artificial intelligence learning technology to improve cutting efficiency and reduce resource waste; it can precisely control the cutting effect, and through intelligent monitoring, timely detect and correct problems in the cutting process to improve the accuracy and stability of cutting; it can provide an adaptive adjustment mechanism, perform self-perception learning according to factors such as the material, yarn count, gram weight, and thickness of the fabric, and score the effect after each cutting. Through continuous evaluation and comparison, the best matching value is selected to determine the most ideal cutting effect, enabling the laser cutting technology to adapt to fabrics of various different materials and characteristics, improving the flexibility and adaptability of cutting; it can save resources, automatically adjust the spot size, height of the laser head, and the running speed of the cutting bed head according to the best matching value, ensuring the accuracy and efficiency of the cutting process, while minimizing fabric waste and defects to achieve resource conservation and utilization; it can improve quality and efficiency, and the cut fabric pieces have advantages such as neat edges and no wire drawing, enhancing the overall aesthetic and quality of the clothing; it can reduce costs, by improving the accuracy and efficiency of the cutting process, reducing the influence of human factors, lowering labor costs and time costs, thereby reducing production costs and enhancing the competitiveness of the enterprise.

[0102] Specifically, the intelligent laser cutting control system for the clothing industry includes a fabric analysis unit, a fabric information identification and reading unit, a cutting judgment unit, a data collection unit, a data cleaning unit, a data set optimization unit, a data preprocessing unit, a laser cutting bed rate regression model training unit, a data collection optimization unit, a lens lighting control unit, an edge-end station cutting picture data collection unit, a laser cutting bed rate regression model deployment unit, a regression unit, and a cutting control unit.

[0103] The fabric analysis unit is used to analyze fabric information. The fabric information recognition and reading unit is used to perform fabric recognition, determine whether the feeding is correct, and when the feeding is correct, read the fabric information. The cutting judgment unit judges whether the cutting is successful. The data acquisition unit is used to perform data acquisition, and the cutting data is transmitted to the server side in real time. The data cleaning unit is used to perform data cleaning. The data set optimization unit optimizes the data set. The data preprocessing unit is used to preprocess the data set optimized by the data set optimization unit. The laser cutting machine speed regression model training unit is used to perform the training of the laser cutting machine speed regression model. The data acquisition optimization unit is used to perform acquisition optimization based on the data of correlation analysis, and optimize the data acquisition content based on the correlation analysis result. The lens and light control unit is used to control the selection of industrial lenses and lights, and control the installation and debugging of industrial lenses and light auxiliary modules. The edge-side station cutting picture data acquisition unit is used to perform the acquisition of edge-side station cutting picture data, and the data is transmitted to the server in real time. The laser cutting machine speed regression model deployment unit is used to perform the deployment of the laser cutting machine parameter control model. The regression unit is used to run the regression model on the cloud platform or the edge side, and perform regression on the optimal cutting speed according to the currently collected real-time fabric data. The cutting control unit is used to perform cutting based on the regressed speed.

[0104] Preferably, the fabric analysis unit is communicatively connected to the cutting management system to analyze the fabric information from the cutting management system.

[0105] More specifically, the fabric information recognition and reading unit includes a fabric recognition module, a feeding judgment module, and a fabric information reading module. The fabric recognition module performs fabric recognition. The feeding judgment module judges whether the feeding is correct. When the feeding is correct, the fabric information reading module reads the fabric information and the laser power p of the laser cutting machine. i . Among them, the fabric information includes the fabric material m i , the fabric density d i , the fabric type (such as yarn count, knitting, weaving, etc.), the gram weight g i , the hardness h i , the flexibility f i , the elasticity e i , etc.

[0106] For example, the fabric information recognition and reading unit is further configured to perform: feeding the fabric based on a barcode scanner, scanning the fabric QR code, and verifying the fabric identification code.

[0107] More specifically, the cutting judgment unit includes a cutting judgment module and an indicator variable module. The cutting judgment module obtains the actually collected cutting speed v i data, and the sewing width q iBased on the data, determine whether the cropping is successful. The indication variable module outputs a cropping success indication variable c i When the cropping is successful, it outputs c i c = 1 indicates that the cropping is successful. When the cropping is not successful, it outputs c i c = 0 indicates that the cropping is not successful.

[0108] Preferably, the data acquisition unit transmits the cropping data to the server side in real time based on the industrial communication protocol.

[0109] More specifically, the data acquisition unit includes a scene judgment module and a weight assignment module. When the scene judgment module determines that the scene is a scene where cropping fails during fabric change, the weight assignment module performs: data weighting in this scene, increasing the weight domain assignment by 4. When the scene judgment module determines that the scene is a scene where cropping starts after several debugging operations are completed after fabric change, the weight assignment module performs: data weighting 5 in this scene. When the scene judgment module determines that the scene is a successful cropping during fabric change, but the change in the tailor width exceeds 20% or more, the weight assignment module performs: data weighting 4 in this scene. When the scene judgment module determines that the scene is a successful cropping during fabric change, but the change in the tailor width is within 20%, the weight assignment module performs: data weighting 5 in this scene. When the scene judgment module determines that the scene is a scene without fabric change, the weight assignment module performs: data weighting 1. When the scene judgment module determines that the scene is a scene where cropping is not successful during the debugging process after fabric change, the weight assignment module performs: data weighting 4 in this scene.

[0110] More specifically, the data cleaning unit includes a missing value processing module, a duplicate data processing module, an error data correction module, a data standardization and normalization module, a format conversion module, an outlier processing module, a data verification module, and a skewness processing module. The missing value processing module performs missing value processing, the duplicate data processing module performs duplicate data processing, the error data correction module performs error data correction, the data standardization and normalization module performs data standardization and normalization, the format conversion module performs format conversion, the outlier processing module performs outlier processing, the data verification module performs data verification, and the skewness processing module performs skewness processing.

[0111] More specifically, the dataset optimization unit includes a filtered weight dataset processing module and a rate optimization dataset processing module. The filtered weight dataset processing module filters all data with a weight assignment of 5 to obtain a weight dataset S c, output to the rate optimization dataset processing module. The rate optimization dataset processing module groups the data according to fabric material, density, type, and gram weight, for example, divides it into 10 groups, obtains the maximum rate of each group of data, then selects the maximum from the maximum rates of each group of data obtained, checks the number of duplicate data entries. If it exceeds the preset threshold, it is considered that the data is not noise, and the rates of all data in this group are modified to the current rate value to obtain a new rate optimization dataset S n 。

[0112] More specifically, the data preprocessing unit is configured to perform: Obtain the rate optimization dataset S n Perform data preprocessing, add a success data field to all data, set the value of the success data item to 1 for all data with a weight assignment of 5, and set the value of the success data item to 0 for data with a weight assignment other than 5.

[0113] More specifically, the data preprocessing unit includes a data integration module, a data transformation module, a feature engineering module, a data binning module, a class balance processing module, and a data skewness processing module. The data integration module performs data integration, the data transformation module performs data transformation, the feature engineering module performs feature engineering preprocessing, the data binning module performs data binning, the class balance processing performs class balance processing, and the data skewness processing module performs data skewness processing.

[0114] Preferably, the training method of the laser cutting machine rate regression model training unit uses XGBoost; wherein, a rate prediction model is trained, and this model can regress the optimal cutting rate based on the information currently faced.

[0115] Preferably, the parameter set sampled by the data acquisition optimization unit is reduced to CH i The parameters included.

[0116] More specifically, the lens lighting control unit includes a lens control module and a lighting control module. The lens control module controls the selection of industrial lenses and controls the installation and debugging of industrial lenses. The lighting control module controls the selection of lighting and controls the installation and debugging of the lighting auxiliary module.

[0117] Preferably, the edge - end station cutting picture data acquisition unit transmits the acquired edge - end station cutting picture data to the server based on the industrial communication protocol.

[0118] More specifically, the laser cutting machine rate regression model deployment unit includes a cloud deployment module and an edge - end deployment module. The cloud deployment module performs the cloud deployment of the laser cutting machine parameter control model, and the edge - end deployment module performs the edge - end deployment of the laser cutting machine parameter control model.

[0119] More specifically, during the cutting process, the cutting control unit observes the cutting effect. If it can cut through, it performs data acquisition and model construction; if it cannot cut through, it records the data during the speed adjustment process of the employees at the cutting station, and trains a new model for iteration after obtaining a new data set.

[0120] Furthermore, the intelligent laser cutting control system for the clothing industry further includes a feature engineering unit, and the feature engineering unit includes a heat map analysis module, a monotonic similarity analysis module, an information gain analysis module, and a feature set determination module.

[0121] The heat map analysis module uses a heat map to analyze the correlation between the foregoing fabric material, fabric density, fabric type, gram weight, cutting bed power, cutting speed, cutting width, and the cutting success indication variable. The monotonic similarity analysis module analyzes the monotonic similarity between these parameters and the cutting success indication variable through the Kendall coefficient. The information gain analysis module analyzes the predictability between these parameters and the cutting success indication variable through information gain. The feature set determination module determines the feature set CH used for predicting the cutting success indication variable based on the predictability analysis of the information gain analysis module. i 。

[0122] Furthermore, the intelligent laser cutting control system for the clothing industry is further configured to execute: constructing a cutting parameter model, where the cutting parameter model processes the input and updated production practice experience and outputs the cutting parameters of the required cutting data for the current station at the edge side. The cutting parameters include but are not limited to fabric thickness, cutting power, cutting width, main fabric material, secondary fabric material, and fabric gram weight.

[0123] Furthermore, the intelligent laser cutting control system for the clothing industry further includes a laser cutting bed parameter control model optimization module, and the laser cutting bed parameter control model optimization is configured to execute: performing data acquisition; loading the laser cutting bed parameter control model; running the laser cutting bed parameter control model; performing speed optimization; setting the optimized speed; observing the cutting effect; determining whether it can cut through; if the determination result is that it cannot cut through, then perform debugging, and after debugging, perform data acquisition to form a new data set after debugging; if the determination result is that it can cut through, then perform data acquisition, construct a new data set, and perform model optimization based on the new data set.

[0124] Those skilled in the art will understand that the present invention has been described with reference to the flowcharts and / or block diagrams of methods, systems, and computer program products according to the present invention. Each block in the flowchart and / or block diagram, as well as combinations of blocks in the flowchart and / or block diagram, can clearly be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, such that the instructions (through the processor of the computer or other programmable data processing device) produce a means for implementing the functions specified in one or more blocks of the flowchart and / or block diagram.

[0125] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and described in the embodiments, and without departing from this principle, the embodiments of the present invention can have any variations or modifications.

Claims

1. An intelligent laser cutting control method for the clothing industry, characterized in that, The intelligent laser cutting control method for the clothing industry includes the following steps: Analyze fabric information; Perform fabric recognition to determine whether the feeding is correct. When the feeding is correct, read the fabric information and the laser power of the laser cutting bed; Collect actual cutting speed data and cutting width data, and determine whether the cutting is successful; Perform data collection, and the cutting data is transmitted to the server side in real time; Perform data cleaning; Optimize the data set, filter out the data with the largest weight assignment to obtain a weighted data set, and process the weighted data set: group the data according to fabric material, density, type, and gram weight, and obtain the maximum rate of each group of data; check the number of duplicate data. If it exceeds the preset threshold, it is determined that the data is not noise, and the rate of all data in this group is modified to the current rate value to obtain an optimized rate optimization data set; Perform data preprocessing on the obtained rate optimization data set; Perform training of the laser cutting bed rate regression model; Perform acquisition optimization based on the data of correlation analysis, and optimize the data acquisition content based on the correlation analysis results; Collect cutting picture data of the edge-side workstations, and the data is transmitted to the server side in real time; Deploy the laser cutting bed parameter control model; Run the regression model on the cloud platform or the edge side, and perform regression on the optimal cutting rate based on the currently collected real-time fabric data; and Perform cutting based on the rate after regression.

2. The intelligent laser cutting control method for the clothing industry according to claim 1, wherein the intelligent laser cutting control method for the clothing industry further includes the steps of: controlling the selection of industrial lenses and lighting, and controlling the installation and debugging of the industrial lens and lighting auxiliary modules.

3. The intelligent laser cutting control method for the clothing industry according to claim 1, wherein the training method for performing the training of the laser cutting bed rate regression model uses XGBoost.

4. The intelligent laser cutting control method for the clothing industry according to claim 1, wherein the steps for performing data cleaning are selected from performing missing value processing, performing duplicate data processing, performing error data correction, performing data standardization, performing data normalization, performing format conversion, performing outlier processing, performing data verification, and performing skewness processing.

5. The intelligent laser cutting control method for the clothing industry according to claim 1, wherein the intelligent laser cutting control method for the clothing industry further includes the steps of: performing data preprocessing on the rate optimization data set, adding a success data field to all data, setting the value of the success data item to 1 for the data with the largest weight assignment, and setting the value of the success data item to 0 for the data with a weight assignment that is not the largest.

6. The intelligent laser cutting control method for the clothing industry as claimed in claim 1, wherein the intelligent laser cutting control method for the clothing industry further comprises the steps of: observing the cutting effect during the cutting process, determining whether the cutting is completed; if the cutting is completed, performing data acquisition and constructing a laser cutting bed parameter control model; if the cutting is not completed, retrieving the data recorded in the rate debugging process, and training the laser cutting bed parameter control model based on the new data set for iteration.

7. The intelligent laser cutting control method for the clothing industry as claimed in any one of claims 1 to 6, wherein the intelligent laser cutting control method for the clothing industry further comprises the steps of: performing data preprocessing selected from performing data integration, performing data transformation, performing feature engineering, performing data binning, performing class balancing processing, and performing data skewness processing.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it executes the steps of the intelligent laser cutting control method for the clothing industry as claimed in any one of claims 1 to 7.

9. An intelligent laser cutting control device for the clothing industry, characterized in that, The intelligent laser cutting control device for the clothing industry comprises: a memory for storing software application programs, a processor for executing the software application programs, and each program of the software application programs correspondingly executes each step of the intelligent laser cutting control method for the clothing industry as claimed in any one of claims 1 to 7.

10. An intelligent laser cutting control system for the clothing industry, which applies the intelligent laser cutting control method for the clothing industry as described in any one of claims 1 to 7, is characterized in that, The intelligent laser cutting control system for the clothing industry includes a fabric analysis unit, a fabric information identification and reading unit, a cutting judgment unit, a data acquisition unit, a data cleaning unit, a dataset optimization unit, a data preprocessing unit, a laser cutting bed rate regression model training unit, a data acquisition optimization unit, a lens lighting control unit, an edge - end station cutting picture data acquisition unit, a laser cutting bed rate regression model deployment unit, a regression unit, and a cutting control unit. Among them, the fabric analysis unit is used to analyze fabric information; the fabric information identification and reading unit is used to perform fabric identification, judge whether the feeding is correct, and when the feeding is correct, read the fabric information; the cutting judgment unit is used to judge whether the cutting is successful; the data acquisition unit is used to perform data acquisition, and the cutting data is transmitted to the server - side in real - time; the data cleaning unit is used to perform data cleaning; the dataset optimization unit optimizes the dataset; the data preprocessing unit is used to preprocess the dataset optimized by the dataset optimization unit; the laser cutting bed rate regression model training unit is used to perform the training of the laser cutting bed rate regression model; the data acquisition optimization unit is used to perform acquisition optimization based on correlation - analysis data, and optimize the data acquisition content based on the correlation - analysis result; the lens lighting control unit is used to control the selection of industrial lenses and lighting, and control the installation and debugging of industrial lenses and lighting auxiliary modules; the edge - end station cutting picture data acquisition unit is used to perform the acquisition of edge - end station cutting picture data, and the data is transmitted to the server in real - time; the laser cutting bed rate regression model deployment unit is used to perform the deployment of the laser cutting bed parameter control model; the regression unit is used to run the regression model on the cloud platform or the edge - end, and perform regression on the optimal cutting rate according to the currently collected real - time fabric data; the cutting control unit is used to perform cutting based on the regressed rate.

11. The intelligent laser cutting control system for the clothing industry according to claim 10, wherein the cutting judgment unit includes a cutting judgment module and an indicator variable module. The cutting judgment module obtains the actually collected cutting speed data and the tailor width data, and judges whether the cutting is successful. The indicator variable module outputs a cutting success indicator variable. When the cutting is successful, it outputs a cutting success indicator variable = 1, indicating that the cutting is successful; when the cutting is not successful, it outputs a cutting success indicator variable = 0, indicating that the cutting is not successful.

12. The intelligent laser cutting control system for the clothing industry according to claim 11, wherein the intelligent laser cutting control system for the clothing industry further includes a feature engineering unit for performing feature engineering.

13. The intelligent laser cutting control system for the clothing industry according to claim 12, wherein the feature engineering unit includes a heat - map analysis module. The heat - map analysis module uses a heat - map to analyze the correlation between fabric material, fabric density, fabric type, gram weight, cutting bed power, cutting speed, tailor width, and the cutting success indicator variable.

14. The intelligent laser cutting control system for the clothing industry according to claim 12, wherein the feature engineering unit includes a monotonic similarity analysis module, and the monotonic similarity analysis module analyzes the monotonic similarity between fabric parameters and the cutting success indication variable through the Kendall coefficient.

15. The intelligent laser cutting control system for the clothing industry according to any one of claims 12 to 14, wherein the feature engineering unit includes an information gain analysis module and a feature set determination module. The information gain analysis module analyzes the predictability between fabric parameters and the cutting success indication variable through information gain, and the feature set determination module determines the feature set used for predicting the cutting success indication variable based on the predictability analysis of the information gain analysis module.

Images