Automatic generation method and system for a garment sewing process

Abstract

The application provides an automatic generation method and system of a garment sewing process, and relates to the technical field of garment design and manufacturing. Firstly, a three-dimensional garment model is generated by using computer-aided technology. Secondly, the three-dimensional model is divided into process units, the connection relationship of the pattern pieces is established, and the initial process flow is determined by cluster analysis. Then, based on dynamic simulation and collision detection algorithms, the fit degree of the process units and the human body is calculated, special structure units are identified, and for the special structure units, a rule engine is applied for process matching, the required special process nodes are inserted into the initial flow, and the insertion position is optimized according to the connection relationship of the pattern pieces. Finally, by constructing a sewing path network, the path length before and after insertion is quantitatively analyzed, and the automatic optimization and adjustment of the process flow are realized. The application realizes the automatic generation and optimization of the garment sewing process based on three-dimensional model and path calculation.

Description

Technical Field

[0001] This invention relates to the field of clothing design and manufacturing technology, specifically to an automatic generation method and system for clothing sewing processes. Background Technology

[0002] The apparel industry is facing the dual challenges of rapidly growing demand for personalized customization and improving production efficiency. Traditional garment sewing process design relies heavily on experienced craftsmen to manually arrange the process, which results in unscientific workflows, low efficiency, and significant fluctuations in finished product quality. With the development of digital manufacturing and intelligent technologies, the automatic generation of garment sewing processes has become an important direction for improving production flexibility and quality control.

[0003] While existing technologies include methods for simulating and optimizing sewing processes based on computer-aided design, digital twins, and discrete event simulation, most of these methods focus on single objectives such as efficiency improvement or quality assurance. They lack systematic solutions for optimizing process sequence, resource coordination, and multiple indicators. Furthermore, there are still challenges in automatically dividing complex garment structures and matching special structural processes, making it difficult to meet the demand for rapid response to diverse styles.

[0004] Therefore, it is necessary to provide an automatic generation method and system for garment sewing processes to solve the aforementioned problem.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an automatic generation method and system for garment sewing processes to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: A method for automatically generating garment sewing processes, comprising the following steps: Step 1: Collect the style design parameters, garment structure dimensions, and fabric parameters of the target garment, and use computer-aided technology to generate a three-dimensional garment model; Step 2: Divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment. Step 3: Dynamically simulate the 3D clothing model based on collision detection and automatic fitting algorithms, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit. Step 4: Based on the rule engine, perform process matching for special structural units, insert the special process nodes required by the special structural units into the preliminary process flow in Step 2, and determine the insertion position according to the connection relationship of the cut pieces; Step 5: Based on the connection relationship between the fabric pieces in the process, construct the sewing path network, calculate the shortest path length between the fabric piece sets of the process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

[0008] Furthermore, the method used to generate three-dimensional clothing models using computer-aided technology is as follows: The collected design parameters, structural dimensions, and fabric parameters of the target garment are imported into the CAD system in a standard format. The two-dimensional pattern diagrams are digitized and the pattern pieces are categorized. A fabric auxiliary attribute database is established. Using the pattern modeling tool in the CAD system, the two-dimensional pattern diagrams are converted into three-dimensional pattern structures. Spatial parameters, including thickness, curvature, and seam allowance, are set for each pattern piece. The seams of each pattern piece are connected according to the garment structure relationship using the "seam" command to form a preliminary three-dimensional model of the target garment. In the preliminary 3D model of the target garment, corresponding fabric parameters are assigned to each piece. A physical simulation algorithm is used to simulate the natural draping, bending, and wrinkling physical effects of the fabric in 3D space. The garment model is then rendered and colored using CAD software to form a 3D garment model.

[0009] Furthermore, the generated 3D garment model is divided into process units to establish the connection relationship between each piece of fabric in each process unit. The method used is as follows: Based on CAD software, all independent cut pieces and their seams in the 3D clothing model are identified. For each seam, its connection type, length and structural location are recorded. The feature vector of each seam is extracted according to the clothing structure and fabric parameters. For each cut piece, the process feature vector of each seam is determined and the process feature vector of each cut piece is summarized. Using cut pieces as nodes and seam edges as edges, a cut piece connection diagram is constructed to represent the connection relationships and seam types between cut pieces. Edge weights between cut pieces are defined, and the physical connection weight values ​​of seam edges are calculated by comprehensively considering seam strength, spatial proximity, and process similarity. The logic behind this is as follows: Two adjacent pieces connected by the seam are defined as a piece pair. The seam strength, spatial proximity and process similarity of each piece pair in the piece connection diagram are processed to be dimensionless. First, the original index data of the piece pairs are collected and integrated into a set of seam strength, a set of centroid distance of pieces and a set of process similarity. Iterate through each set and select the maximum value of the center-seam strength of the fabric pieces. and minimum value Maximum value between centroids and minimum value and the maximum value of process similarity and minimum value First, dimensionless processing is performed, based on the following formula: in, , , These represent the dimensionless suture strength index, spatial proximity index, and process similarity index, respectively. Indicates the first The individual cut piece and the first The stitching strength between individual fabric pieces Indicates the first The individual cut piece and the first The similarity of the processes between individual cut pieces For the first The individual cut piece and the first The distance between the centroids of the individual cut pieces , The index of the uniformly cut pieces, and , , This refers to the total number of cut pieces; The physical connection weight of the fabric piece pair is calculated based on the dimensionless seam strength, spatial proximity, and process similarity of the seam edges. The formula used is as follows: in, Indicates the first The individual cut piece and the first The physical connection weight value between individual cut pieces , , These represent the weighting coefficients for the suture strength index, spatial proximity index, and process similarity index, respectively. For each fabricated piece's process feature vector, a Gaussian kernel function is used to calculate the feature similarity between pairs of pieces. This similarity is then adjusted by combining the comprehensive edge weights of the seams between the pieces. The formula used is as follows: in, Indicates the first The individual cut piece and the first Feature similarity values ​​between individual cut pieces , The first The cut piece and the first The process feature vector of each cut piece, This is a bandwidth parameter used to control the rate at which similarity decays. Indicates the first The individual cut piece and the first The final adjacency weight values ​​between each piece of fabric; An adjacency matrix is ​​constructed based on the calculated adjacency weight values ​​between the cut pieces. The graph Laplacian matrix is ​​constructed using spectral clustering analysis, based on the following formula: in, Indicates the first The sum of the connection strengths of each piece reflects the first... The cumulative size of the edge weights between this piece and all other pieces. Represent a The diagonal matrix represents the sum of the adjacency weights of all the cut pieces. The symbol for generating a diagonal matrix. It is a graph Laplacian matrix used to represent the overall connection relationship between pattern pieces.

[0010] Furthermore, cluster analysis was used to determine the specific processes for cutting the garment pieces and to establish the preliminary workflow for the garment. The method used was as follows: Obtain the Graph Laplace Matrix Calculate the eigenvalues ​​and corresponding eigenvectors of . The former The eigenvectors corresponding to the smallest eigenvalues ​​form a matrix. ,in This represents the preset number of clusters, determined based on production requirements and process unit size constraints, for the matrix. Each row is normalized by dividing the vector in each row by its Euclidean norm to obtain the normalized feature matrix. K-value cluster analysis was used to... Clustering is performed on the row vectors, with each row corresponding to a pattern piece. The clustering result is the division of the pattern pieces, divided into: Each process unit ; Statistical analysis is performed on the process feature vectors of all cut pieces within the same process unit. The mean, variance, maximum and minimum values ​​of each feature are calculated. Representative process parameters, including fabric thickness, fabric elasticity range, cut piece type distribution, and sewing complexity level, are extracted. Based on the statistical results and garment structure information, the fabric type, structural parts, and process complexity level of the process unit are mapped to a set of multi-dimensional labels. Each label dimension corresponds to a key process attribute, which is used to describe the process characteristics and processing requirements of the process unit. Using the multi-dimensional label set, combined with the sewing edge weight and process dependency, the specific sewing process scheme of the process unit is defined, including the sewing sequence, sewing parameters, and process priority.

[0011] Furthermore, dynamic simulations are performed on the 3D clothing model to calculate the fit between each process unit and the human body, in order to identify special structural units in the 3D clothing model. The method used is as follows: Based on the 3D clothing model divided into process units, coordinate alignment, scale normalization, and mesh optimization are performed to ensure accurate spatial correspondence between the clothing and the human body model. Based on collision detection principles, the system detects whether there is interpenetration or overlap between the clothing and the human body. Accelerated collision detection is used to check the closest distance and overlap between the clothing mesh and the human body mesh frame by frame. All vertices or faces with interpenetration or obvious gaps are marked, and the distribution and severity statistics of collision points between each process unit and the human body are output. The uniformity of the fit is assessed using a fabric simulation method to fit dynamic fall data, based on the following formula: in, This indicates the fit of each process unit. This indicates the first process unit in the process unit. The distance from each sampling point to the human body surface, This is the index of the sampling point in the process unit, and , This represents the total number of sampling points in the process unit. It is the variance of the fit of the process unit, used to represent the uniformity of the fit of the process unit; A uniformity threshold is established, and the fit variance of each process unit is compared with the threshold. The process unit containing the fit variance exceeding the threshold is defined as a special structural unit in the three-dimensional clothing model.

[0012] Furthermore, process matching is performed on the special structural units, inserting the special process nodes required by the special structural units into the preliminary process flow in step 2. The method used is as follows: For each type of special structural unit, process rules are established. The rules include process type, required process nodes, and process parameters. The structural label and process feature vector of each labeled structural unit are used as input to the rule engine. The corresponding process rules are automatically retrieved based on the structural label, the newly added special process nodes and their parameters are determined, and the list of special process nodes and their attributes are output. Let the set of cut pieces be The basic process nodes are Each process node is associated with a subset of cut pieces, and similarly, each special process node is also associated with a subset of cut pieces. For each special process node, the system searches for the basic process node with the closest correlation to its cut pieces in the preliminary process flow diagram. For each special process node that needs to be inserted, the system searches for the basic process node with the largest intersection with its cut piece set in the preliminary process flow. After determining the reference basic process node, the insertion method is determined according to the nature of the special process node: if the special process node belongs to the preprocessing of the cut pieces, it is inserted before the reference process node; if it belongs to the subsequent finishing processing, it is inserted after the process node. When inserting in the process flow diagram, the dependencies between the original nodes are adjusted. For the case of insertion before, all connections between the preceding processes pointing to the reference process node and that node need to be disconnected and changed to point to the newly inserted special process node. Then, the special process node points to the reference process node. For insertion after, the reference process node points to the newly inserted special process node, and the subsequent process nodes that the reference process node originally pointed to are now pointed to by the newly inserted node.

[0013] Furthermore, a sewing path network is constructed, and the shortest path length between the sets of cut pieces at the nodes of the preceding and following processes is calculated to determine whether the insertion position needs to be readjusted. The method used is as follows: Treat each garment piece as a node and the stitching relationships between pieces as undirected edges, constructing a sewing path network graph. ,in Represents a set of cut pieces. Let be the set of seam connections between the fabric pieces, and let the set of fabric pieces corresponding to the special process node to be inserted be . The set of cut pieces corresponding to the process node before the insertion position is The set of cut pieces corresponding to the process node after the insertion position is The formula used to calculate the shortest path length between the sets of cut pieces from the preceding and following process nodes before insertion is: in, This represents the shortest path distance between sets of cut pieces from preceding and following processes, without passing through any special process nodes. , They respectively represent belonging to the set of cut pieces , A specific cut piece node represents a cut piece corresponding to a process node before and after the insertion position, respectively. Cutting node To the cutting point The shortest path length between the pieces is calculated based on the edge weights between the pieces. After calculating the path length after inserting the special process node, which requires passing through the set of cut pieces corresponding to the special process, the path length is: in, This indicates that after inserting a special process node, the path must pass through the set of cut pieces corresponding to the special process, and the shortest path length is calculated. This represents the set of cut pieces corresponding to a special process node. A specific cut piece node represents a cut piece involved in a special structural unit. For the previous process cutting node To the special process cutting node The shortest path length, To cut pieces from special process nodes To the subsequent cutting process node The shortest path length; Obtained through calculation , Let tolerance coefficient be set. If the conditions are met If the insertion position is correct, it is considered reasonable. If this condition is not met, it needs to be readjusted.

[0014] The present invention also provides an automatic generation system for garment sewing processes, the automatic generation system being used to execute the above-described automatic generation method for garment sewing processes, comprising: The 3D clothing modeling module is used to collect the style design parameters, clothing structure dimensions, and fabric parameters of the target clothing, and to generate a 3D clothing model using computer-aided technology. The process unit division module is used to divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment. The dynamic fit simulation module uses collision detection and automatic fitting algorithms to dynamically simulate the 3D clothing model, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit. The special process insertion rule engine module performs process matching on special structural units based on the rule engine, inserts the special process nodes required by the special structural units into the preliminary process flow in step 2, and determines the insertion position based on the connection relationship of the cut pieces. The process rationality verification module is used to construct a sewing path network based on the connection relationship between the cut pieces in the process, calculate the shortest path length between the cut piece sets of process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

[0015] Compared with the prior art, the beneficial effects of the present invention are: Based on technologies such as computer-aided design, 3D garment modeling, cluster analysis, collision detection, and rule engine, this invention realizes the automatic division and process generation of garment sewing processes. By constructing a pattern connection diagram and using a spectral clustering algorithm, process units are scientifically divided. Combined with dynamic fit simulation, special structural units are accurately identified, and corresponding special process nodes are automatically matched and inserted. This solves the problems of traditional process design relying on experience, unscientific processes, and difficulty in handling special structures, and realizes the systematization and intelligence of process design. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall method flow of the present invention.

[0017] Figure 2 This is a schematic diagram of the system module flow of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0019] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0020] Example: Please see Figure 1An automatic generation method for garment sewing processes, the specific steps of which include: Step 1: Collect the style design parameters, garment structure dimensions, and fabric parameters of the target garment, and use computer-aided technology to generate a three-dimensional garment model.

[0021] In a specific embodiment of the present invention, by importing the style design parameters, structural dimensions and fabric parameters of the target garment into a CAD system, a highly accurate three-dimensional garment model can be generated. This high-precision model ensures the true reproduction of the design intent, avoids errors caused by manual drawing or simple modeling, and by setting spatial parameters for each piece, the fabric characteristics and process requirements can be accurately reflected in the model, thereby providing basic data for subsequent process division and process optimization.

[0022] Furthermore, the method used to generate three-dimensional clothing models using computer-aided technology is as follows: The collected design parameters, structural dimensions, and fabric parameters of the target garment are imported into the CAD system in a standard format. The two-dimensional pattern diagrams are digitized and the pattern pieces are categorized. A fabric auxiliary attribute database is established. Using the pattern modeling tool in the CAD system, the two-dimensional pattern diagrams are converted into three-dimensional pattern structures. Spatial parameters, including thickness, curvature, and seam allowance, are set for each pattern piece. The seams of each pattern piece are connected according to the garment structure relationship using the "seam" command to form a preliminary three-dimensional model of the target garment. In the preliminary 3D model of the target garment, corresponding fabric parameters are assigned to each piece. A physical simulation algorithm is used to simulate the natural draping, bending, and wrinkling physical effects of the fabric in 3D space. The garment model is then rendered and colored using CAD software to form a 3D garment model.

[0023] Step 2: Divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment.

[0024] In a specific embodiment of the present invention, by dividing the three-dimensional garment model into detailed process units, establishing the connection relationship between each piece and calculating its physical connection weight, and using Gaussian kernel function and spectral clustering analysis, a piece connection diagram and a graph Laplacian matrix are generated, which can accurately characterize the connection relationship and sewing type between pieces. This method makes the process characteristics and connection relationship between pieces more systematic and quantitative.

[0025] It should be noted that in the formula for constructing physical connection weights, the sewing strength reflects the physical connection strength between the cut pieces, taking into account the influence of fabric characteristics and sewing methods on the connection firmness. Spatial proximity is used to quantify spatial relationships by calculating the distance between the centroids of the cut pieces. Closer cut pieces usually have a higher synergistic effect when sewing. Therefore, the edge weight coefficient is used to adjust the distance influence and improve the rationality of the connection. Process similarity is based on the similarity of the process characteristics of the cut pieces. Cut pieces with high process similarity usually have better sewing results when connected, reflecting the design consistency between the cut pieces. By defining nodes and edges, a clear pattern connection diagram is formed, which allows the structural relationships of clothing design to be presented intuitively. The constructed graph Laplacian matrix provides a foundation for subsequent spectral clustering analysis, which can effectively identify and classify pattern pieces with similar features.

[0026] It should be noted that, based on the target 3D garment model, the process unit is formed by rationally grouping multiple pieces of fabric in the entire garment model according to sewing techniques, structural relationships, and process characteristics, creating relatively independent sewing units that are easy to manage and execute. Using a CAD system or 3D modeling software, all pieces of fabric are numbered and their attributes are labeled, the seam boundaries between pieces are identified, and the type, length, position, and structural part of each seam are recorded. For each seam, the sewing strength, spatial proximity, and seam process characteristics are extracted. The process characteristics of each piece containing seams are summarized to form a piece-level process feature vector. Furthermore, a process unit contains multiple pieces of fabric that are sewn together, and each piece belongs to one and only one process unit to achieve full coverage and mutual exclusion.

[0027] Furthermore, the generated 3D garment model is divided into process units to establish the connection relationship between each piece of fabric in each process unit. The method used is as follows: Based on CAD software, all independent cut pieces and their seams in the 3D clothing model are identified. For each seam, its connection type, length and structural location are recorded. The feature vector of each seam is extracted according to the clothing structure and fabric parameters. For each cut piece, the process feature vector of each seam is determined and the process feature vector of each cut piece is summarized. Using cut pieces as nodes and seam edges as edges, a cut piece connection diagram is constructed to represent the connection relationships and seam types between cut pieces. Edge weights between cut pieces are defined, and the physical connection weight values ​​of seam edges are calculated by comprehensively considering seam strength, spatial proximity, and process similarity. The logic behind this is as follows: Two adjacent pieces connected by the seam are defined as a piece pair. The seam strength, spatial proximity and process similarity of each piece pair in the piece connection diagram are processed to be dimensionless. First, the original index data of the piece pairs are collected and integrated into a set of seam strength, a set of centroid distance of pieces and a set of process similarity. Iterate through each set and select the maximum value of the center-seam strength of the fabric pieces. and minimum value Maximum value between centroids and minimum value and the maximum value of process similarity and minimum value First, dimensionless processing is performed, based on the following formula: in, , , These represent the dimensionless suture strength index, spatial proximity index, and process similarity index, respectively. Indicates the first The individual cut piece and the first The stitching strength between individual fabric pieces Indicates the first The individual cut piece and the first The similarity of the processes between individual cut pieces For the first The individual cut piece and the first The distance between the centroids of the individual cut pieces , The index of the uniformly cut pieces, and , , This refers to the total number of cut pieces; The physical connection weight of the fabric piece pair is calculated based on the dimensionless seam strength, spatial proximity, and process similarity of the seam edges. The formula used is as follows: in, Indicates the first The individual cut piece and the first The physical connection weight value between individual cut pieces , , These represent the weighting coefficients for the suture strength index, spatial proximity index, and process similarity index, respectively. For each fabricated piece's process feature vector, a Gaussian kernel function is used to calculate the feature similarity between pairs of pieces. This similarity is then adjusted by combining the comprehensive edge weights of the seams between the pieces. The formula used is as follows: in, Indicates the first The individual cut piece and the first Feature similarity values ​​between individual cut pieces , The first The cut piece and the first The process feature vector of each cut piece, This is a bandwidth parameter used to control the rate at which similarity decays. Indicates the first The individual cut piece and the first The final adjacency weight values ​​between each piece of fabric; An adjacency matrix is ​​constructed based on the calculated adjacency weight values ​​between the cut pieces. The graph Laplacian matrix is ​​constructed using spectral clustering analysis, based on the following formula: in, Indicates the first The sum of the connection strengths of each piece reflects the first... The cumulative size of the edge weights between this piece and all other pieces. Represent a The diagonal matrix represents the sum of the adjacency weights of all the cut pieces. The symbol for generating a diagonal matrix. It is a graph Laplacian matrix used to represent the overall connection relationship between pattern pieces.

[0028] It should be noted that cluster analysis is used to divide the fabric pieces into process segments and determine the preliminary process flow. This method utilizes the eigenvalues ​​and eigenvectors of the graph Laplacian matrix to extract and normalize the feature matrix, thereby achieving efficient clustering of the fabric pieces. The clustering results provide a clear division for each process unit. Then, the process characteristics of the fabric pieces within the same process unit are statistically analyzed to extract key process parameters and map them into multidimensional labels. This process not only improves the systematization and standardization of process characteristics, but also optimizes the sewing sequence, sewing parameters, and process priority by defining specific sewing process schemes, thereby significantly improving the efficiency and quality of the production process.

[0029] Furthermore, cluster analysis was used to determine the specific processes for cutting the garment pieces and to establish the preliminary workflow for the garment. The method used was as follows: Obtain the Graph Laplace Matrix Calculate the eigenvalues ​​and corresponding eigenvectors of . The former The eigenvectors corresponding to the smallest eigenvalues ​​form a matrix. ,in This represents the preset number of clusters, determined based on production requirements and process unit size constraints, for the matrix. Each row is normalized by dividing the vector in each row by its Euclidean norm to obtain the normalized feature matrix. K-value cluster analysis was used to... Clustering is performed on the row vectors, with each row corresponding to a pattern piece. The clustering result is the division of the pattern pieces, divided into: Each process unit ; Statistical analysis is performed on the process feature vectors of all cut pieces within the same process unit. The mean, variance, maximum and minimum values ​​of each feature are calculated. Representative process parameters, including fabric thickness, fabric elasticity range, cut piece type distribution, and sewing complexity level, are extracted. Based on the statistical results and garment structure information, the fabric type, structural parts, and process complexity level of the process unit are mapped to a set of multi-dimensional labels. Each label dimension corresponds to a key process attribute, which is used to describe the process characteristics and processing requirements of the process unit. Using the multi-dimensional label set, combined with the sewing edge weight and process dependency, the specific sewing process scheme of the process unit is defined, including the sewing sequence, sewing parameters, and process priority.

[0030] Step 3: Dynamically simulate the 3D clothing model based on collision detection and automatic fitting algorithms, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit.

[0031] In a specific embodiment of the present invention, the fit between the 3D clothing model and the human body is calculated through dynamic simulation. This method can effectively identify special structural units. First, by using coordinate alignment, scale normalization, and mesh optimization, the accurate spatial correspondence between the clothing and the human body model is ensured. Second, by using the collision detection principle, the distance and overlap between the clothing mesh and the human body mesh are detected frame by frame, and areas with interlacing or obvious gaps are accurately marked. Finally, the fit and fit variance of each process unit are calculated by using the fabric simulation method, and the process units are classified using a uniformity threshold, effectively identifying special structural units.

[0032] The innovation of using the distance from sampling points to the human body surface to quantify fit lies in its direct correlation between the adaptability of the 3D clothing model and the human body model. By evaluating the distance of each sampling point, the fit between the clothing and the human body can be captured more meticulously. This method not only improves the measurement accuracy of fit, but also quantifies and analyzes the relative positional relationship between the clothing and the human body at different parts, thereby identifying potential design problems or areas for improvement.

[0033] Furthermore, dynamic simulations are performed on the 3D clothing model to calculate the fit between each process unit and the human body, in order to identify special structural units in the 3D clothing model. The method used is as follows: Based on the 3D clothing model divided into process units, coordinate alignment, scale normalization, and mesh optimization are performed to ensure accurate spatial correspondence between the clothing and the human body model. Based on collision detection principles, the system detects whether there is interpenetration or overlap between the clothing and the human body. Accelerated collision detection is used to check the closest distance and overlap between the clothing mesh and the human body mesh frame by frame. All vertices or faces with interpenetration or obvious gaps are marked, and the distribution and severity statistics of collision points between each process unit and the human body are output. The uniformity of the fit is assessed using a fabric simulation method to fit dynamic fall data, based on the following formula: in, This indicates the fit of each process unit. This indicates the first process unit in the process unit. The distance from each sampling point to the human body surface, This is the index of the sampling point in the process unit, and , This represents the total number of sampling points in the process unit. It is the variance of the fit of the process unit, used to represent the uniformity of the fit of the process unit; A uniformity threshold is set, and the fit variance of the process unit is compared with the threshold. The process unit where the fit variance exceeds the threshold is defined as a special structural unit in the three-dimensional clothing model. It should be noted that special structural units refer to process units where the fit variance exceeds a preset threshold. These process units are composed of several pieces connected together, representing special parts of the garment, such as shoulder pleats, cuff joints, and complex structural areas like the collar curve. These areas have complex shapes, large fabric deformations, or significant changes in the curvature of the human body, making the sewing process more complex and requiring the insertion of special process nodes for targeted processing.

[0034] Step 4: Based on the rule engine, perform process matching for special structural units, insert the special process nodes required by the special structural units into the preliminary process flow in Step 2, and determine the insertion position according to the connection relationship of the cut pieces.

[0035] In a specific embodiment of this invention, by matching special structural units with processes and inserting them into the initial process flow, this method significantly improves the flexibility and adaptability of the process flow. Specifically, by establishing process rules and automatically retrieving corresponding rules, customized process nodes and parameter configurations are provided for each special structural unit. This matching and insertion mechanism ensures the integrity of the process flow while optimizing the processing of each special structural unit, thereby improving overall production efficiency and product quality. By dynamically adjusting process nodes and their dependencies, the system can quickly respond to design changes and production needs, enhancing the flexibility and intelligence level of the production line.

[0036] The specific method for formulating logical rules for inserting special structural units into the process flow is as follows: First, process rule matching is performed. Specific process processing rules are established for each type of special structural unit, including process type, required process nodes, and process parameters. Structural labels and process feature vectors are used as inputs to the rule engine to automatically retrieve and match the corresponding process rules, and determine the special process nodes and their parameters that need to be added.

[0037] The logic for inserting special structural units is as follows: In the preliminary process flow diagram, find the basic process node with the closest correlation to the special process node by the cut piece set. This is determined by the intersection of the cut piece sets and divided into two types. The first type is a preprocessing node, which is a special process node that is a preprocessing step of the cut piece. In this case, it is inserted before the reference process node. This requires adjusting all the preceding processes that point to the reference process node so that they point to the newly inserted special process node. The second type is a subsequent processing node. If the special process node is a subsequent finishing process, it is inserted after the reference process node. The original connection pointing to the subsequent process node is changed to be pointed to by the newly inserted node.

[0038] Furthermore, process matching is performed on the special structural units, inserting the special process nodes required by the special structural units into the preliminary process flow in step 2. The method used is as follows: For each type of special structural unit, process rules are established. The rules include process type, required process nodes, and process parameters. The structural label and process feature vector of each labeled structural unit are used as input to the rule engine. The corresponding process rules are automatically retrieved based on the structural label, the newly added special process nodes and their parameters are determined, and the list of special process nodes and their attributes are output. Let the set of cut pieces be The basic process nodes are Each process node is associated with a subset of cut pieces, and similarly, each special process node is also associated with a subset of cut pieces. For each special process node, the system searches for the basic process node with the closest correlation to its cut pieces in the preliminary process flow diagram. For each special process node that needs to be inserted, the system searches for the basic process node with the largest intersection with its cut piece set in the preliminary process flow. After determining the reference basic process node, the insertion method is determined according to the nature of the special process node: if the special process node belongs to the preprocessing of the cut pieces, it is inserted before the reference process node; if it belongs to the subsequent finishing processing, it is inserted after the process node. When inserting in the process flow diagram, the dependencies between the original nodes are adjusted. For the case of insertion before, all connections between the preceding processes pointing to the reference process node and that node need to be disconnected and changed to point to the newly inserted special process node. Then, the special process node points to the reference process node. For insertion after, the reference process node points to the newly inserted special process node, and the subsequent process nodes that the reference process node originally pointed to are now pointed to by the newly inserted node.

[0039] Step 5: Based on the connection relationship between the fabric pieces in the process, construct the sewing path network, calculate the shortest path length between the fabric piece sets of the process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

[0040] In a specific embodiment of this invention, by constructing a sewing path network and calculating the shortest path length between the sets of fabric pieces before and after the insertion of a special process node, a quantitative judgment on the rationality of the insertion position of the special process node is achieved. This method can effectively avoid problems such as abnormal growth of the sewing path and discontinuity of the process flow caused by the insertion of special processes, thereby ensuring the overall efficiency and rationality of the garment sewing process.

[0041] The reason for setting up a step to verify the rationality of the sewing process is that in the actual garment sewing process, special processes often need to be inserted into specific positions in the main sewing process. If the insertion position is not chosen properly, it may lead to a significant increase in the original sewing path, affecting the production cycle, efficiency and finished product quality. Therefore, by using network modeling and quantitative analysis of path length, we can scientifically and systematically determine whether the insertion position is appropriate, and avoid process deviations caused by experience or subjective judgment.

[0042] Furthermore, a sewing path network is constructed, and the shortest path length between the sets of cut pieces at the nodes of the preceding and following processes is calculated to determine whether the insertion position needs to be readjusted. The method used is as follows: Treat each garment piece as a node and the stitching relationships between pieces as undirected edges, constructing a sewing path network graph. ,in Represents a set of cut pieces. Let be the set of seam connections between the fabric pieces, and let the set of fabric pieces corresponding to the special process node to be inserted be . The set of cut pieces corresponding to the process node before the insertion position is The set of cut pieces corresponding to the process node after the insertion position is The formula used to calculate the shortest path length between the sets of cut pieces from the preceding and following process nodes before insertion is: in, This represents the shortest path distance between sets of cut pieces from preceding and following processes, without passing through any special process nodes. , They respectively represent belonging to the set of cut pieces , A specific cut piece node represents a cut piece corresponding to a process node before and after the insertion position, respectively. Cutting node To the cutting point The shortest path length between the pieces is calculated based on the edge weights between the pieces. After calculating the path length after inserting the special process node, which requires passing through the set of cut pieces corresponding to the special process, the path length is: in, This indicates that after inserting a special process node, the path must pass through the set of cut pieces corresponding to the special process, and the shortest path length is calculated. This represents the set of cut pieces corresponding to a special process node. A specific cut piece node represents a cut piece involved in a special structural unit. For the previous process cutting node To the special process cutting node The shortest path length, To cut pieces from special process nodes To the subsequent cutting process node The shortest path length; Obtained through calculation , Let tolerance coefficient be set. If the conditions are met If the insertion position is correct, it is considered reasonable. If this condition is not met, it needs to be readjusted.

[0043] Please see Figure 2 The present invention also provides an automatic generation system for garment sewing processes, the automatic generation system being used to execute the above-mentioned automatic generation method for garment sewing processes, comprising: The 3D clothing modeling module is used to collect the style design parameters, clothing structure dimensions, and fabric parameters of the target clothing, and to generate a 3D clothing model using computer-aided technology. The process unit division module is used to divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment. The dynamic fit simulation module uses collision detection and automatic fitting algorithms to dynamically simulate the 3D clothing model, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit. The special process insertion rule engine module performs process matching on special structural units based on the rule engine, inserts the special process nodes required by the special structural units into the preliminary process flow in step 2, and determines the insertion position based on the connection relationship of the cut pieces. The process rationality verification module is used to construct a sewing path network based on the connection relationship between the cut pieces in the process, calculate the shortest path length between the cut piece sets of process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

[0044] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0045] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0046] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for automatically generating garment sewing processes, characterized in that, The specific steps include: Step 1: Collect the style design parameters, garment structure dimensions, and fabric parameters of the target garment, and use computer-aided technology to generate a three-dimensional garment model; Step 2: Divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment. Step 3: Dynamically simulate the 3D clothing model based on collision detection and automatic fitting algorithms, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit. Step 4: Based on the rule engine, perform process matching for special structural units, insert the special process nodes required by the special structural units into the preliminary process flow in Step 2, and determine the insertion position according to the connection relationship of the cut pieces; Step 5: Based on the connection relationship between the fabric pieces in the process, construct the sewing path network, calculate the shortest path length between the fabric piece sets of the process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

2. The automatic generation method for garment sewing processes according to claim 1, characterized in that, The method used to generate 3D clothing models using computer-aided technology is as follows: The collected design parameters, structural dimensions, and fabric parameters of the target garment are imported into the CAD system in a standard format. The two-dimensional pattern diagrams are digitized and the pattern pieces are categorized. A fabric auxiliary attribute database is established. Using the pattern modeling tool in the CAD system, the two-dimensional pattern diagrams are converted into three-dimensional pattern structures. Spatial parameters, including thickness, curvature, and seam allowance, are set for each pattern piece. The seams of each pattern piece are connected according to the garment structure relationship using the "sewing" command to form a preliminary three-dimensional model of the target garment. In the preliminary 3D model of the target garment, corresponding fabric parameters are assigned to each piece. A physical simulation algorithm is used to simulate the natural draping, bending, and wrinkling physical effects of the fabric in 3D space. The garment model is then rendered and colored using CAD software to form a 3D garment model.

3. The automatic generation method for garment sewing processes according to claim 2, characterized in that, The generated 3D garment model is divided into process units to establish the connection relationship between each piece of fabric in each process unit. The method used is as follows: Based on CAD software, all independent cut pieces and their seams in the 3D clothing model are identified. For each seam, its connection type, length and structural location are recorded. The feature vector of each seam is extracted according to the clothing structure and fabric parameters. For each cut piece, the process feature vector of each seam is determined and the process feature vector of each cut piece is summarized. Using cut pieces as nodes and seam edges as edges, a cut piece connection diagram is constructed to represent the connection relationships and seam types between cut pieces. Edge weights between cut pieces are defined, and the physical connection weight values ​​of seam edges are calculated by comprehensively considering seam strength, spatial proximity, and process similarity. The logic behind this is as follows: Two adjacent pieces connected by the seam are defined as a piece pair. The seam strength, spatial proximity and process similarity of each piece pair in the piece connection diagram are processed to be dimensionless. First, the original index data of the piece pairs are collected and integrated into a set of seam strength, a set of centroid distance of pieces and a set of process similarity. Iterate through each set and select the maximum value of the center stitch strength of the cut pieces. and minimum value Maximum value between centroids and minimum value and the maximum value of process similarity and minimum value First, dimensionless processing is performed, based on the following formula: in, , , These represent the dimensionless suture strength index, spatial proximity index, and process similarity index, respectively. Indicates the first The individual cut piece and the first The stitching strength between individual fabric pieces Indicates the first The individual cut piece and the first The similarity of the processes between individual cut pieces For the first The individual cut piece and the first The distance between the centroids of the individual cut pieces , The index of the uniformly cut pieces, and , , This refers to the total number of cut pieces; The physical connection weight of the fabric piece pair is calculated based on the dimensionless seam strength, spatial proximity, and process similarity of the seam edges. The formula used is as follows: in, Indicates the first The individual cut piece and the first The physical connection weight value between individual cut pieces , , These represent the weighting coefficients for the suture strength index, spatial proximity index, and process similarity index, respectively. For each fabricated piece's process feature vector, a Gaussian kernel function is used to calculate the feature similarity between pairs of pieces. This similarity is then adjusted by combining the comprehensive edge weights of the seams between the pieces. The formula used is as follows: in, Indicates the first The individual cut piece and the first Feature similarity values ​​between individual cut pieces , The first The cut piece and the first The process feature vector of each cut piece, This is a bandwidth parameter used to control the rate at which similarity decays. Indicates the first The individual cut piece and the first The final adjacency weight values ​​between each piece of fabric; An adjacency matrix is ​​constructed based on the calculated adjacency weight values ​​between the cut pieces. The graph Laplacian matrix is ​​constructed using spectral clustering analysis, based on the following formula: in, Indicates the first The sum of the connection strengths of each piece reflects the first... The cumulative size of the edge weights between this piece and all other pieces. Represent a The diagonal matrix represents the sum of the adjacency weights of all the cut pieces. The symbol for generating a diagonal matrix. It is a graph Laplacian matrix used to represent the overall connection relationship between pattern pieces.

4. The automatic generation method for garment sewing processes according to claim 3, characterized in that, Cluster analysis was used to determine the specific process for cutting the garment pieces and to establish the preliminary workflow for the garment. The method used was as follows: Obtain the Graph Laplacian Matrix Calculate the eigenvalues ​​and corresponding eigenvectors of . The former The eigenvectors corresponding to the smallest eigenvalues ​​form a matrix. ,in This represents the preset number of clusters, determined based on production requirements and process unit size constraints, for the matrix. Each row is normalized by dividing the vector in each row by its Euclidean norm to obtain the normalized feature matrix. K-value cluster analysis was used to... Clustering is performed on the row vectors, with each row corresponding to a pattern piece. The clustering result is the division of the pattern pieces, divided into: Each process unit ; Statistical analysis is performed on the process feature vectors of all cut pieces within the same process unit. The mean, variance, maximum and minimum values ​​of each feature are calculated. Representative process parameters, including fabric thickness, fabric elasticity range, cut piece type distribution, and sewing complexity level, are extracted. Based on the statistical results and garment structure information, the fabric type, structural parts, and process complexity level of the process unit are mapped to a set of multi-dimensional labels. Each label dimension corresponds to a key process attribute, which is used to describe the process characteristics and processing requirements of the process unit. Using the multi-dimensional label set, combined with the sewing edge weight and process dependency, the specific sewing process scheme of the process unit is defined, including the sewing sequence, sewing parameters, and process priority.

5. The automatic generation method for garment sewing processes according to claim 4, characterized in that, Dynamic simulation of a 3D clothing model is performed to calculate the fit between each process unit and the human body in order to identify special structural units in the 3D clothing model. The method used is as follows: Based on the 3D clothing model divided into process units, coordinate alignment, scale normalization, and mesh optimization are performed to ensure accurate spatial correspondence between the clothing and the human body model. Based on collision detection principles, the system detects whether there is interpenetration or overlap between the clothing and the human body. Accelerated collision detection is used to check the closest distance and overlap between the clothing mesh and the human body mesh frame by frame. All vertices or faces with interpenetration or obvious gaps are marked, and the distribution and severity statistics of collision points between each process unit and the human body are output. The uniformity of the fit is assessed using a fabric simulation method to fit dynamic fall data, based on the following formula: in, This indicates the fit of each process unit. This indicates the first process unit in the process unit. The distance from each sampling point to the human body surface, This is the index of the sampling point in the process unit, and , This represents the total number of sampling points in the process unit. It is the variance of the fit of the process unit, used to represent the uniformity of the fit of the process unit; A uniformity threshold is established, and the fit variance of each process unit is compared with the threshold. The process unit with a fit variance exceeding the threshold is defined as a special structural unit in the three-dimensional clothing model.

6. The method for automatically generating garment sewing processes according to claim 5, characterized in that, The process matching for special structural units involves inserting the specific process nodes required by these units into the preliminary process flow in step 2. The method used is as follows: For each type of special structural unit, process rules are established. The rules include process type, required process nodes, and process parameters. The structural label and process feature vector of each labeled structural unit are used as input to the rule engine. The corresponding process rules are automatically retrieved based on the structural label, the newly added special process nodes and their parameters are determined, and the list of special process nodes and their attributes are output. Let the set of cut pieces be The basic process nodes are Each process node is associated with a subset of cut pieces, and similarly, each special process node is also associated with a subset of cut pieces. For each special process node, the system searches for the basic process node with the closest correlation to its cut pieces in the preliminary process flow diagram. For each special process node that needs to be inserted, the system searches for the basic process node with the largest intersection with its cut piece set in the preliminary process flow. After determining the reference basic process node, the insertion method is determined according to the nature of the special process node: if the special process node belongs to the preprocessing of the cut pieces, it is inserted before the reference process node; if it belongs to the subsequent finishing processing, it is inserted after the process node. When inserting in the process flow diagram, the dependencies between the original nodes are adjusted. For the case of insertion before, all connections between the preceding processes pointing to the reference process node and that node need to be disconnected and changed to point to the newly inserted special process node. Then, the special process node points to the reference process node. For insertion after, the reference process node points to the newly inserted special process node, and the subsequent process nodes that the reference process node originally pointed to are now pointed to by the newly inserted node.

7. The method for automatically generating garment sewing processes according to claim 6, characterized in that, A sewing path network is constructed, and the shortest path length between the sets of cut pieces at the nodes of the preceding and following processes is calculated to determine whether the insertion position needs to be readjusted. The method used is as follows: Treat each garment piece as a node and the stitching relationships between pieces as undirected edges, constructing a sewing path network graph. ,in Represents a set of cut pieces. Let be the set of seam connections between the fabric pieces, and let the set of fabric pieces corresponding to the special process node to be inserted be . The set of cut pieces corresponding to the process node before the insertion position is The set of cut pieces corresponding to the process node after the insertion position is The formula used to calculate the shortest path length between the sets of cut pieces from the preceding and following process nodes before insertion is: in, This represents the shortest path distance between sets of cut pieces from preceding and following processes, without passing through any special process nodes. , They respectively represent belonging to the set of cut pieces , A specific cut piece node represents a cut piece corresponding to a process node before and after the insertion position, respectively. Cutting node To the cutting point The shortest path length between the pieces is calculated based on the edge weights between the pieces. After calculating the path length after inserting the special process node, which requires passing through the set of cut pieces corresponding to the special process, the path length is: in, This indicates that after inserting a special process node, the path must pass through the set of cut pieces corresponding to the special process, and the shortest path length is calculated. This represents the set of cut pieces corresponding to a special process node. A specific cut piece node represents a cut piece involved in a special structural unit. For the previous process cutting node To the special process cutting node The shortest path length, To cut pieces from special process nodes To the subsequent cutting process node The shortest path length; Obtained based on calculation , Let the tolerance coefficient be set. If the conditions are met If the insertion position is correct, it is considered reasonable. If this condition is not met, it needs to be readjusted.

8. An automatic generation system for garment sewing processes, characterized in that, The automatic generation system is used to execute the automatic generation method for garment sewing processes according to any one of claims 1-7, including: The 3D clothing modeling module is used to collect the style design parameters, clothing structure dimensions, and fabric parameters of the target clothing, and to generate a 3D clothing model using computer-aided technology. The process unit division module is used to divide the generated 3D garment model into process units, establish the connection relationship between each piece of fabric in each process unit, and use cluster analysis to determine the specific process of the fabric pieces in order to determine the preliminary process flow of the garment. The dynamic fit simulation module uses collision detection and automatic fitting algorithms to dynamically simulate the 3D clothing model, calculate the fit between each process unit and the human body, and identify special structural units in the 3D clothing model based on the uniformity of the fit. The special process insertion rule engine module performs process matching on special structural units based on the rule engine, inserts the special process nodes required by the special structural units into the preliminary process flow in step 2, and determines the insertion position based on the connection relationship of the cut pieces. The process rationality verification module is used to construct a sewing path network based on the connection relationship between the cut pieces in the process, calculate the shortest path length between the cut piece sets of process nodes before and after insertion, and compare the path after insertion with the direct path before insertion to determine whether the insertion position needs to be readjusted.

Images