Garment pattern making method and system based on human body three-dimensional model

Abstract

The invention discloses a clothing pattern making method and system based on a human body three-dimensional model. The method comprises the following steps: firstly, acquiring human body three-dimensional data information through a high-precision human body scanner, preprocessing the human body three-dimensional data information, converting the human body three-dimensional data information into human body three-dimensional point cloud data, and constructing a human body 3D model according to the human body three-dimensional point cloud data; performing classification and feature point extraction on human body 3D model data by combining geometric features of human body parts through an AI model; a human body surface three-dimensional grid is generated based on the generated feature points, the human body surface three-dimensional grid is refined according to a preset three-dimensional data structure and a rapid capture algorithm, and the refined three-dimensional surface is expanded to obtain a two-dimensional surface patch; finally, the unfolded pieces are spliced to obtain a human body two-dimensional prototype, and a clothing template is manufactured. According to the invention, complex three-dimensional human body information can be converted into various two-dimensional prototypes which contain all body shape data and can be directly used for clothing pattern making.

Description

Technical Field

[0001] The present invention relates to the technical field of clothing pattern making, and in particular to a clothing pattern making method and system based on a three-dimensional human body model. Background Art

[0002] Pattern making technology is a key link in clothing customization, which determines the efficiency, fit, comfort and aesthetic feeling of clothing customization. In the Internet era, the demand for clothing customization is increasing day by day, and consumers' customization needs are complex.

[0003] Traditional customization technology mainly uses body measurements, but does not have specific and accurate body shape data, such as chest height angle, hunchback angle, etc., resulting in a high return rate for finished garments. The overall pattern-making technology is relatively backward, and technological development mainly relies on the traditional master-apprentice teaching method, which requires years of experience and exploration, and lacks the support of digital scientific theoretical basis. Summary of the invention

[0004] Based on this, the embodiments of the present application provide a clothing pattern making method and system based on a three-dimensional human body model, which can convert complex three-dimensional human body information into various two-dimensional prototypes containing all body shape data that can be directly used for clothing pattern making, and while achieving automation, high efficiency and accuracy, it can also meet customers' personalized needs for customization.

[0005] In a first aspect, a method for making a clothing pattern based on a three-dimensional human body model is provided, the method comprising:

[0006] Acquire three-dimensional data information of a human body through a high-precision human body scanner, convert the three-dimensional data information of a human body into three-dimensional point cloud data of a human body after preprocessing, and construct a 3D model of a human body according to the three-dimensional point cloud data of a human body;

[0007] Through AI model, human body 3D model data is classified and feature points are extracted in combination with the geometric features of human body parts;

[0008] A three-dimensional mesh of the human body surface is generated based on the generated feature points, the three-dimensional mesh of the human body surface is refined according to a preset three-dimensional data structure and a fast capture algorithm, and the refined three-dimensional surface is unfolded to obtain a two-dimensional surface patch;

[0009] The unfolded pieces are spliced ​​together to obtain a two-dimensional prototype of the human body and to make a clothing template.

[0010] Optionally, after constructing a 3D human body model according to the 3D point cloud data of the human body, the method further comprises:

[0011] The human body 3D model data is processed through a graph detection scheme, noise and error data are deleted, and the human body 3D model data is uploaded to a database; wherein the noise and error data may be redundant triangular meshes.

[0012] Optionally, the human body 3D model data is classified and feature points are extracted by combining the geometric features of human body parts through the AI ​​model, including:

[0013] The geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

[0014] Optionally, the refined three-dimensional surface is unfolded to obtain a two-dimensional surface patch, including:

[0015] Extract the surface area to be covered from the refined three-dimensional surface; divide the surface area into small blocks, each of which contains multiple discrete sampling points;

[0016] The kd-tree data structure is used to index these sampling points so that the nearest neighbor points can be quickly queried; for the points that need to generate a grid on the surface, the kd-tree search algorithm is used to find the nearest neighbor points that fall on the surface of the human body model.

[0017] These nearest neighbor points are used to construct small-sized triangular meshes, and all small-sized triangular meshes are spliced ​​to obtain a two-dimensional patch.

[0018] Optionally, the step of splicing the unfolded pieces to obtain a two-dimensional human body prototype comprises:

[0019] The various parts on the three-dimensional model are manually unfolded into graphics on a two-dimensional plane according to certain rules; the unfolded parts are spliced ​​according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, the two-dimensional prototype is adjusted according to actual conditions, the adjusted two-dimensional prototype is digitized and output into a computer-readable format.

[0020] In a second aspect, a clothing pattern making system based on a three-dimensional human body model is provided, the system comprising:

[0021] A preprocessing module is used to obtain three-dimensional data information of a human body through a high-precision human body scanner, convert the three-dimensional data information of the human body into three-dimensional point cloud data of the human body after preprocessing, and construct a 3D model of the human body according to the three-dimensional point cloud data of the human body;

[0022] The extraction module is used to classify the human body 3D model data and extract feature points by combining the AI ​​model with the geometric features of human body parts;

[0023] An unfolding module is used to generate a three-dimensional mesh of the human body surface based on the generated feature points, refine the three-dimensional mesh of the human body surface according to a preset three-dimensional data structure and a fast capture algorithm, and unfold the refined three-dimensional surface to obtain a two-dimensional surface patch;

[0024] The splicing module is used to splice the unfolded pieces to obtain a two-dimensional prototype of the human body and make a clothing template.

[0025] Optionally, after the preprocessing module constructs a 3D human body model according to the human body 3D point cloud data, it further includes:

[0026] The human body 3D model data is processed through a graph detection scheme, noise and error data are deleted, and the human body 3D model data is uploaded to a database; wherein the noise and error data may be redundant triangular meshes.

[0027] Optionally, the extraction module includes:

[0028] The geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

[0029] Optionally, the expansion module includes:

[0030] Extract the surface area to be covered from the refined three-dimensional surface; divide the surface area into small blocks, each of which contains multiple discrete sampling points;

[0031] The kd-tree data structure is used to index these sampling points so that the nearest neighbor points can be quickly queried; for the points that need to generate a grid on the surface, the kd-tree search algorithm is used to find the nearest neighbor points that fall on the surface of the human body model.

[0032] These nearest neighbor points are used to construct small-sized triangular meshes, and all small-sized triangular meshes are spliced ​​to obtain a two-dimensional patch.

[0033] Optionally, the splicing module includes:

[0034] The various parts on the three-dimensional model are manually unfolded into graphics on a two-dimensional plane according to certain rules; the unfolded parts are spliced ​​according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, the two-dimensional prototype is adjusted according to actual conditions, the adjusted two-dimensional prototype is digitized and output into a computer-readable format.

[0035] In the technical solution provided in the embodiment of the present application, the three-dimensional data information of the human body is first obtained through a high-precision human body scanner, and the three-dimensional data information of the human body is converted into three-dimensional point cloud data after pre-processing, and a 3D model of the human body is constructed based on the three-dimensional point cloud data of the human body; then, the 3D model data of the human body is classified and feature points are extracted through the AI ​​model in combination with the geometric features of the human body parts; the three-dimensional mesh of the human body surface is generated based on the generated feature points, and the three-dimensional mesh of the human body surface is refined according to the preset three-dimensional data structure and fast capture algorithm, and the refined three-dimensional surface is unfolded to obtain a two-dimensional surface piece; finally, the unfolded pieces are spliced ​​to obtain a two-dimensional prototype of the human body and a clothing template is made. It can be seen that the beneficial effects of the present invention include at least:

[0036] (1) Use high-precision scanning to obtain the customer's 3D body shape data, achieve data losslessness and standardization, and facilitate customer data storage, transmission, and analysis.

[0037] (2) The human 3D body model data is automatically uploaded to the database.

[0038] (3) Providing human prototypes for clothing pattern making. Based on human body characteristics, the scanned 3D human body model is automatically converted into a 2D prototype that contains all customer body shape data and can be directly used for clothing pattern making. This realizes the conversion of human body data from 3D to 2D, provides data basis and theoretical support for personalized clothing pattern making, and solves the problem of ill-fitting customization.

[0039] (4) Establish a customer body shape data storage system to store customer body shape data in the most efficient way, which can be retrieved at any time, providing big data support for enterprises to develop new clothing and patterns. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0041] Figure 1 A flow chart of a clothing pattern making method based on a three-dimensional human body model provided in an embodiment of the present application;

[0042] Figure 2 A block diagram of a clothing pattern making system based on a three-dimensional human body model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0044] In the description of the present invention, the terms "comprises", "has" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or apparatus comprising a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may also include other steps or units that are not explicitly listed but are inherent to these processes, methods, products or apparatuses, or steps or units added based on further optimization schemes conceived by the present invention.

[0045] The traditional body measurement method is too simple, and the body measurement cannot accurately express the complex body shape of people, which is the key reason why the custom-made clothes do not fit. For example: the traditional custom shops use body measurements, but the plate making technology, style, research and development, etc. cannot meet customer needs; the custom factories use body measurements, but there is no body shape data, and the pattern cannot match the customer's body shape, resulting in a high repair rate for the ready-made clothes; brand companies use body measurements, but there is no body shape data, and they cannot effectively store and analyze and perform fitting operations. They need to adjust the pattern samples many times to achieve the purpose of fit, resulting in a continuous increase in production costs and repair rates.

[0046] This technology directly unfolds the scanned three-dimensional human body model into a two-dimensional plane, and automatically generates a two-dimensional plate-making prototype (three-section body, four-section body, knife back, etc.) containing all human body shape data, so that the human body and clothing can be reasonably combined, providing a data basis and theoretical support for personalized clothing customization, solving the problem of ill-fitting patterns, improving customization efficiency, reducing customization costs, and greatly improving the quality of clothing in all aspects.

[0047] Specifically, please refer to Figure 1 , which shows a flowchart of a clothing pattern making method based on a three-dimensional human body model provided in an embodiment of the present application. The method may include the following steps:

[0048] Step 101, obtaining three-dimensional data information of a human body through a high-precision human body scanner, converting the three-dimensional data information of the human body into three-dimensional point cloud data of the human body after preprocessing, and constructing a 3D model of the human body according to the three-dimensional point cloud data of the human body.

[0049] In this step, a high-precision human body scanner is used to automatically scan and obtain three-dimensional data information of the human body;

[0050] Specifically, high-precision human body scanners can include laser scanning, RGB-D cameras, structured light scanning, etc., and use high-precision human body scanners to scan the human body and obtain data of the human body 3D model. This data can include information such as body size, curves, proportions, and internal structure information such as muscles, organs, and bones. This data will become the basis for subsequent design and production processes.

[0051] The human body 3D point cloud data is converted into a triangular mesh to build a human body 3D model, and the human body 3D model data is automatically uploaded to the database. The human body 3D model data is processed to delete redundant triangular meshes; a fast graph detection solution is specifically used to process the model, delete noise and error data, and obtain a more accurate 3D model; at the same time, the 3D model can also be compressed with low loss, which has almost no impact on accuracy while accelerating processing.

[0052] In this embodiment, the specific process of deleting redundant triangular meshes includes:

[0053] Finding redundant triangles: First, we need to determine which triangles are redundant. We can determine whether each triangle is adjacent to the surrounding triangles by calculating the normal vector of each triangle. If the normal vector of a triangle is the same as the normal vector of its adjacent triangles, it can be considered a redundant triangle.

[0054] Removing redundant triangles: Once the redundant triangles are identified, they need to be removed from the mesh. This can be done by deleting their vertices from the vertex list. After deleting the vertices, the triangles and edges associated with them must be updated to maintain the integrity of the model.

[0055] Remesh: After removing redundant triangles, holes or broken edges may remain. These problems can be fixed by remeshing. The specific method of remeshing depends on the needs of the application, but generally involves adding new triangles or quads to fill holes and adjusting existing vertices and edges to ensure the mesh topology is correct.

[0056] Mesh optimization: To further reduce redundancy and improve mesh quality, the mesh can be optimized. This may include removing over-tessellated triangles, smoothing triangle surfaces to eliminate aliasing effects, or reprojecting texture coordinates to eliminate distortion.

[0057] Step 102, classifying the human body 3D model data and extracting feature points by combining the geometric features of human body parts through the AI ​​model.

[0058] In this step, the geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

[0059] The curvature, two-dimensional directional gradient and inflection point calculation of the three-dimensional human body model can be achieved by the following steps:

[0060] Calculation of 3D curvature: 3D curvature is a metric that describes the variation in surface shape. Conventional methods (e.g., based on Gaussian or mean curvature measurements) can be used to calculate the curvature value for each vertex. This allows detection of areas with high curvature (e.g., finger joints and spine) and low curvature (e.g., arms and legs) in different locations.

[0061] 2D Directional Gradient Sum Calculation: 2D Directional Gradient Sum is a technique commonly used in image processing and can also be applied to 3D models. For each surface point, its normal vector and the tangent vector along the direction of maximum and minimum curvature at that point can be calculated. Then, the angle between these two tangent vectors can be calculated and the result can be mapped to a range of 0 to 1. This gives a value that represents the local geometric features. Areas with high Directional Gradient Sum values ​​may be areas with significant changes such as corners or edges.

[0062] Inflection point detection: Inflection points are points on a surface where the direction of curvature changes abruptly. Inflection points can be detected by calculating the first and second order derivatives of the curvature. Regions with a large number of inflection points may be very detailed and complex regions on the surface, such as fingers and facial features.

[0063] By calculating the curvature and 2D directional gradients and inflection points of the 3D human body model, different geometric features can be detected and important feature points can be extracted from them. These feature points can be used for applications such as classification, posture estimation and action recognition.

[0064] Step 103, based on the three-dimensional mesh of the human body surface generated by the generated feature points, the three-dimensional mesh of the human body surface is refined according to a preset three-dimensional data structure and a fast capture algorithm, and the refined three-dimensional surface is unfolded to obtain a two-dimensional surface patch.

[0065] In this step, a three-dimensional mesh of the human body surface based on feature points is automatically generated, and the generated three-dimensional mesh of the human body surface is subdivided piece by piece based on triangle strips; through a suitable three-dimensional data structure and a fast capture algorithm, the three-dimensional mesh is quickly refined, thereby reducing the errors generated in the process of clothing prototype generation.

[0066] The specific process of refining the three-dimensional grid is as follows:

[0067] Create a three-dimensional data structure: Usually use a data structure such as an octree or a quadtree to divide the three-dimensional space into equal-sized sub-blocks and store the corresponding grid points in each sub-block.

[0068] Load original mesh: Load the original 3D mesh into a data structure and generate a unique identifier for each mesh point for subsequent processing.

[0069] Snapping algorithm: Get a certain number of sampling points from the original mesh using a fast snapping algorithm (such as octree-basedpointsampling). These sampling points can be equally spaced or feature-based. The higher the density of sampling points, the higher the quality of the generated fine mesh.

[0070] Generate a new mesh: Use the sampling points and the topological information of the original mesh to generate a new, more refined 3D mesh. There are several methods:

[0071] Interpolation based on sampling points: For each sampling point, the information of surrounding grid points is used for interpolation to infer the new grid point position. The interpolation method can be linear interpolation, cubic interpolation, etc.

[0072] Surface reconstruction: Generate a surface model based on the sampling points and their normal vector information. This method can better preserve the details of the original model.

[0073] Data fitting: The sampling points are fitted by using curve fitting, surface fitting, etc. This method can reconstruct the original mesh more accurately, but it also requires more computing resources.

[0074] Post-processing: After the new 3D mesh is generated, it can be further processed, such as removing overlapping parts, smoothing, etc.

[0075] In this step, a fast three-dimensional surface unfolding scheme is used to unfold the generated fine three-dimensional surface to obtain a two-dimensional surface patch.

[0076] The surface area to be covered is extracted from the refined three-dimensional surface; the surface area is divided into small blocks, each of which contains multiple discrete sampling points;

[0077] The process of using kd-tree for point cloud search is as follows:

[0078] The surface area that needs to be covered is extracted from the 3D human body model.

[0079] The surface area is divided into small blocks, each of which contains some discrete sampling points (point cloud).

[0080] Use the kd-tree data structure to index these sampling points so that the nearest neighbor points can be quickly queried. Kd-tree is a binary tree structure with a partitioning hyperplane at each node, which can classify all data points according to a certain dimension, so that the leaf node closest to the query point can be quickly found on the tree.

[0081] For the points that need to generate a grid on the surface, the nearest neighbor points that fall on the surface of the human body model are found through the kd-tree search algorithm.

[0082] Use these nearest neighbor points to construct a small-sized triangular mesh. Specifically, one of the points can be connected to several of its nearest neighbor points first, and then based on these nearest neighbor points, the boundary segments between them are connected until a series of small-sized triangular meshes are formed.

[0083] Finally, by splicing all the small-sized triangular meshes together, we can get a detailed human body model surface unfolding result, that is, a two-dimensional patch.

[0084] Step 104, splicing the unfolded pieces to obtain a two-dimensional human body prototype and making a clothing template.

[0085] In this step, the unfolded pieces are spliced ​​to obtain a two-dimensional prototype of the human body; this process combines the design experience of multiple clothing pattern designers, and performs translation and rotation operations on the pieces to make the obtained clothing prototype sample more reasonable and more suitable for later design and use.

[0086] Specifically, the parts on the three-dimensional model are manually unfolded into graphics on a two-dimensional plane according to certain rules; the unfolded parts are spliced ​​according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, and the two-dimensional prototype is adjusted according to actual conditions, and the adjusted two-dimensional prototype is digitized and output into a computer-readable format. A customer body shape data storage system is established to store customer body shape data in the most efficient way.

[0087] In the embodiments of the present application, the whole process is visualized, and manual intervention and adjustment can be performed; the process can be displayed in real time, and convenient interaction can be performed to customize the three-dimensional model of the human body, thereby more fully meeting the personalized needs of customers and designers.

[0088] From the above, it can be seen that the key technical points of the present invention are:

[0089] First, the three-dimensional data information of the human body is acquired through scanning, and a two-dimensional prototype is automatically generated from the three-dimensional human body model. The complex three-dimensional human body information is converted into various two-dimensional prototypes that contain all body shape data and can be directly used for clothing pattern making, providing data and theoretical support for personalized clothing pattern making.

[0090] During this process, the present invention is based on a modern computer processing system, adopts a variety of efficient and practical algorithms, and independently creates and develops algorithms such as model detection and repair, feature extraction, three-dimensional surface segmentation of human body models, rapid three-dimensional surface expansion, and prototype splicing. While achieving automation, high efficiency, and accuracy, it can also meet customers' personalized needs for customization.

[0091] like Figure 2 The embodiment of the present application also provides a clothing pattern making system based on a three-dimensional human body model.

[0092] The system includes:

[0093] A preprocessing module is used to obtain three-dimensional data information of a human body through a high-precision human body scanner, convert the three-dimensional data information of the human body into three-dimensional point cloud data of the human body after preprocessing, and construct a 3D model of the human body according to the three-dimensional point cloud data of the human body;

[0094] The extraction module is used to classify the human body 3D model data and extract feature points by combining the AI ​​model with the geometric features of human body parts;

[0095] An unfolding module is used to generate a three-dimensional mesh of the human body surface based on the generated feature points, refine the three-dimensional mesh of the human body surface according to a preset three-dimensional data structure and a fast capture algorithm, and unfold the refined three-dimensional surface to obtain a two-dimensional surface patch;

[0096] The splicing module is used to splice the unfolded pieces to obtain a two-dimensional prototype of the human body and make a clothing template.

[0097] In an optional embodiment of the present application, after the preprocessing module constructs the human body 3D model according to the human body 3D point cloud data, it also includes:

[0098] The human body 3D model data is processed through a graph detection scheme, noise and error data are deleted, and the human body 3D model data is uploaded to a database; wherein the noise and error data may be redundant triangular meshes.

[0099] In an optional embodiment of the present application, the extraction module includes:

[0100] The geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

[0101] In an optional embodiment of the present application, the expansion module includes:

[0102] Extract the surface area to be covered from the refined three-dimensional surface; divide the surface area into small blocks, each of which contains multiple discrete sampling points;

[0103] The kd-tree data structure is used to index these sampling points so that the nearest neighbor points can be quickly queried; for the points that need to generate a grid on the surface, the kd-tree search algorithm is used to find the nearest neighbor points that fall on the surface of the human body model.

[0104] These nearest neighbor points are used to construct small-sized triangular meshes, and all small-sized triangular meshes are spliced ​​to obtain a two-dimensional patch.

[0105] In an optional embodiment of the present application, the splicing module includes: manually unfolding various parts on the three-dimensional model into graphics on a two-dimensional plane according to certain rules; splicing the unfolded parts according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, adjusting the two-dimensional prototype according to actual conditions, digitizing the adjusted two-dimensional prototype, and outputting it into a computer-readable format.

[0106] The clothing plate making system based on the three-dimensional human body model provided in the embodiment of the present application is used to implement the above-mentioned clothing plate making method based on the three-dimensional human body model. For the specific definition of the clothing plate making system based on the three-dimensional human body model, please refer to the definition of the clothing plate making method based on the three-dimensional human body model in the above text, which will not be repeated here. The various parts of the above-mentioned clothing plate making system based on the three-dimensional human body model can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the device in the form of hardware, or can be stored in the memory in the device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0107] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0108] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the patent application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent application shall be subject to the attached claims.

Claims

1. A clothing pattern making method based on a three-dimensional human body model, characterized in that: The method comprises: Acquire three-dimensional data information of a human body through a high-precision human body scanner, convert the three-dimensional data information of a human body into three-dimensional point cloud data of a human body after preprocessing, and construct a 3D model of a human body according to the three-dimensional point cloud data of a human body; Through AI model, human body 3D model data is classified and feature points are extracted in combination with the geometric features of human body parts; A three-dimensional mesh of the human body surface is generated based on the generated feature points, the three-dimensional mesh of the human body surface is refined according to a preset three-dimensional data structure and a fast capture algorithm, and the refined three-dimensional surface is unfolded to obtain a two-dimensional surface patch; The unfolded pieces are spliced ​​together to obtain a two-dimensional prototype of the human body and to make a clothing template.

2. The method for making a clothing plate according to claim 1, characterized in that: After constructing a 3D human body model according to the 3D point cloud data of the human body, the method further comprises: The human body 3D model data is processed through a graph detection scheme, noise and error data are deleted, and the human body 3D model data is uploaded to a database; wherein the noise and error data may be redundant triangular meshes.

3. The method for making a clothing plate according to claim 1, characterized in that: Through AI models, combined with the geometric features of human body parts, human 3D model data is classified and feature points are extracted, including: The geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

4. The method for making a clothing plate according to claim 1, characterized in that: The refined three-dimensional surface is unfolded to obtain a two-dimensional surface patch, including: Extract the surface area to be covered from the refined three-dimensional surface; divide the surface area into small blocks, each of which contains multiple discrete sampling points; Use the kd-tree data structure to index these sampling points so that the nearest neighbor points can be quickly queried; for the points that need to generate a grid on the surface, the kd-tree search algorithm is used to find the nearest neighbor points that fall on the surface of the human body model; These nearest neighbor points are used to construct small-sized triangular meshes, and all small-sized triangular meshes are spliced ​​to obtain a two-dimensional patch.

5. The method for making a clothing plate according to claim 1, characterized in that: The step of splicing the unfolded pieces to obtain a two-dimensional human body prototype comprises: The various parts on the three-dimensional model are manually unfolded into graphics on a two-dimensional plane according to certain rules; the unfolded parts are spliced ​​according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, the two-dimensional prototype is adjusted according to actual conditions, the adjusted two-dimensional prototype is digitized and output into a computer-readable format.

6. A clothing pattern making system based on a three-dimensional human body model, characterized in that: The system comprises: A preprocessing module is used to obtain three-dimensional data information of a human body through a high-precision human body scanner, convert the three-dimensional data information of the human body into three-dimensional point cloud data of the human body after preprocessing, and construct a 3D model of the human body according to the three-dimensional point cloud data of the human body; The extraction module is used to classify the human body 3D model data and extract feature points by combining the AI ​​model with the geometric features of human body parts; An unfolding module is used to generate a three-dimensional mesh of the human body surface based on the generated feature points, refine the three-dimensional mesh of the human body surface according to a preset three-dimensional data structure and a fast capture algorithm, and unfold the refined three-dimensional surface to obtain a two-dimensional surface patch; The splicing module is used to splice the unfolded pieces to obtain a two-dimensional prototype of the human body and make a clothing template.

7. The clothing pattern making system according to claim 6, characterized in that: After the preprocessing module constructs a 3D human body model according to the 3D point cloud data of the human body, it also includes: The human body 3D model data is processed through a graph detection scheme, noise and error data are deleted, and the human body 3D model data is uploaded to a database; wherein the noise and error data may be redundant triangular meshes.

8. The clothing pattern making system according to claim 6, characterized in that: The extraction module includes: The geometric features of the human body are detected by calculating the curvature, two-dimensional directional gradient and inflection point of the three-dimensional human body model.

9. The clothing pattern making system according to claim 6, characterized in that: The expansion modules include: Extract the surface area to be covered from the refined three-dimensional surface; divide the surface area into small blocks, each of which contains multiple discrete sampling points; Use the kd-tree data structure to index these sampling points so that the nearest neighbor points can be quickly queried; for the points that need to generate a grid on the surface, the kd-tree search algorithm is used to find the nearest neighbor points that fall on the surface of the human body model; These nearest neighbor points are used to construct small-sized triangular meshes, and all small-sized triangular meshes are spliced ​​to obtain a two-dimensional patch.

10. The clothing pattern making system according to claim 6, characterized in that: The splicing module includes: The various parts on the three-dimensional model are manually unfolded into graphics on a two-dimensional plane according to certain rules; the unfolded parts are spliced ​​according to their corresponding relationships on the three-dimensional model to generate a complete two-dimensional prototype of the human body, the two-dimensional prototype is adjusted according to actual conditions, the adjusted two-dimensional prototype is digitized and output into a computer-readable format.

Images