Dynamic self-adaptive garment pattern generation method and system based on soft tissue deformation and multi-posture human body measurement

Abstract

The invention discloses a dynamic self-adaptive garment pattern generation method and system based on soft tissue deformation and multi-posture human body measurement, and belongs to the field of intelligent garment design. The method comprises the steps of multi-posture human body data acquisition, data preprocessing and database construction, soft tissue layering mechanical modeling, dynamic version parameter calculation, version automatic adjustment, cloth simulation verification, iterative optimization and feedback correction. The system comprises a data acquisition module, a human body deformation modeling module, an automatic model optimization module and the like, the modules are connected through a specific communication mode, a layered mechanical model is established by acquiring three-dimensional data and soft tissue deformation data of six typical postures of a human body, and cloth physical simulation and feedback correction are combined, so that the model is optimized. According to the method, automatic optimization of garment cutting pieces, stitches and elastic areas is achieved, the fitness and comfort of the garment in a dynamic posture are ensured, the intelligent and personalized level of garment design is improved, and the method has remarkable practical value and popularization prospects.

Description

Technical Field

[0001] This invention relates to a dynamic adaptive clothing pattern generation method and system. More specifically, it relates to a dynamic adaptive clothing pattern generation method and system that combines soft tissue deformation modeling, three-dimensional human body measurement, and virtual simulation. It belongs to the field of intelligent clothing design and human body modeling technology and is applicable to multiple fields such as personalized customized clothing production, intelligent fitting equipment, ergonomic clothing design, and sportswear research and development. It can realize the intelligent generation of clothing patterns that dynamically adapt to multiple postures of the human body. Background Technology

[0002] Current garment pattern design techniques primarily rely on static human body model data. Designers typically use anthropometric parameters of standard body types or fixed standing postures as the basis for pattern drawing and design. However, in daily life and during exercise, the human body frequently undergoes postural changes such as bending, sitting, walking, raising arms, and twisting. During these processes, soft tissues such as skin, fat, and muscles undergo significant stretching, contraction, and slippage deformations. These dynamic deformations directly affect the fit, comfort, and appearance of clothing. Although existing technologies such as virtual fitting systems and 3D garment simulation tools have emerged to assist in design, simulating the effect of clothing under different body types to some extent, these technologies are mostly limited to static geometric analysis of body shape differences, lacking in-depth modeling of the deformation patterns of human soft tissues and their dynamic responses in multiple postures. While existing technologies can only simulate the fit of clothing in a fixed posture, they cannot accurately predict the strain distribution, wrinkle formation, and slippage of clothing during movement. This leads to problems such as localized tightness, excessive wrinkles, and slippage in actual wear, severely impacting user experience and causing industry pain points such as high return rates and low customization efficiency for apparel companies. Therefore, current technologies are insufficient to meet consumers' demands for dynamic adaptability and personalized comfort in clothing. There is an urgent need for a method and system that can automatically generate clothing patterns that match dynamic postures based on multi-posture anthropometric data and comprehensive soft tissue deformation laws, providing a technological foundation for fundamentally achieving dynamic adaptation and intelligent customization in clothing design. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of existing technologies in clothing pattern design, such as relying on static human body models and ignoring the dynamic deformation of soft tissues, resulting in poor fit and insufficient comfort. It provides a dynamic adaptive clothing pattern generation method and system based on soft tissue deformation and multi-pose human body measurement. By collecting multi-pose 3D data of the human body and soft tissue deformation data, a precise layered mechanical model is established to achieve automatic optimization of clothing pieces, seams, and elastic areas. Combined with fabric physical simulation and feedback correction mechanisms, this ensures that clothing maintains good fit and comfort in various daily postures, ultimately achieving intelligent, personalized, and dynamically adaptable clothing design.

[0004] To achieve the above objectives, the present invention is implemented through the following technical solution: This invention discloses a dynamic adaptive clothing pattern generation method based on soft tissue deformation and multi-pose anthropometrics, comprising the following steps: Step 1, Multi-pose human body data acquisition: Collect 3D point cloud or mesh data of the human body in various typical poses, and at the same time set up flexible wearable sensors at key deformation parts of the human body to collect strain values, slippage and displacement data of the skin surface. Step 2, Data Preprocessing and Database Construction: The multi-pose 3D data collected in Step 1 are registered, and a structured mesh model is generated through topology optimization. The deformation vector field, principal strain direction and strain distribution matrix of the human body surface under each pose are extracted. Combined with human body shape parameters, a multi-dimensional human body deformation database is established. Each sample in the database is associated with a pose category, body shape parameter group and corresponding soft tissue deformation parameter set. Step 3, Soft tissue layered mechanical modeling: Using a multi-layer coupled mechanical model based on finite element analysis, the human soft tissue is divided into muscle layer, fat layer and skin layer. The deformation characteristics of each layer are described by the corresponding mechanical model, and the model parameters are calibrated by sample data in the database. Step 4, Dynamic Pattern Parameter Calculation: Receive the body shape parameters input by the user or the uploaded 3D scan model, combine them with the user's selected daily high-frequency postures, call the soft tissue mechanical model established in Step 3, calculate the surface strain distribution, tensile amount and slip trajectory of each part of the human body under the target posture, and generate a dynamic surface model of the human body. Step 5, Automatic Adjustment of Garment Pattern: Based on the dynamic surface model in Step 4, the corresponding initial pattern library is called according to the garment type. The outline of the cut pieces is automatically corrected through a parametric adjustment algorithm, the seam layout is optimized, and elastic fabric blocks are configured in the corresponding parts. The elastic modulus of the elastic fabric blocks is adaptively adjusted according to the amount of slippage. Step 6, Fabric physical simulation verification: Obtain the physical parameters of the clothing fabric, use a mass spring model to simulate the fabric, simulate the fit, stretching and wrinkle distribution of the clothing under the high-frequency posture selected by the user, and output the fabric stress value of each part of the simulation results. Step 7, Pattern Iteration Optimization: If there are stress concentration areas or areas with excessive bonding gaps in the simulation results, automatically return to Step 5 to adjust the cut piece parameters, seam positions, or elastic fabric configuration until the preset optimization target is met; Step 8, Feedback Correction and Model Update: The fabric strain data during actual wear is collected by the flexible strain sensor built into the clothing, compared with the simulation results of Step 6, the error value is calculated, the parameters of the soft tissue mechanical model are corrected by a neural network algorithm, and the relevant user data is added to the human body deformation database of Step 2 to realize model self-learning. Preferably, the key deformable parts of the human body in step 1 include the shoulder, elbow, waist, hip, knee joint, ankle joint and abdomen. The flexible wearable sensor is a flexible resistive strain sensor with a strain measurement range of 0-50%, a slip measurement accuracy of ±0.1mm, and a displacement data sampling frequency of not less than 100Hz. The typical postures in step 1 include standing, bending over, sitting, walking, raising arms and twisting. When bending over, the forward tilt angle of the human body is 30°-90°. When sitting, the knee bending angle is 90°. When walking, the step frequency is 50-120 steps / minute. When raising arms, the angle between the arm and the torso is 90°-180°. When twisting, the left and right torso twists by 30°-60°.

[0005] Preferably, the registration in step 2 adopts the iterative nearest point algorithm, the registration error is controlled within ±0.3mm, the topology optimization adopts the Poisson reconstruction algorithm, and the human body shape parameters include height, weight, chest circumference, waist circumference, hip circumference, body fat percentage, and muscle mass. The number of samples in the multidimensional human deformity database is not less than 10,000. Preferably, in step 3, the muscle layer uses the Hill muscle model to describe its contraction characteristics, the fat layer uses the Ogden hyperelastic model to describe its nonlinear deformation, and the skin layer uses the linear elastic model to describe its tensile characteristics. The model parameters are calibrated using the least squares method, with the actual deformation data of the samples in the database as the target value. The mechanical parameters of each layer are determined through iterative calculation. The deformation prediction error of the calibrated model does not exceed 5%. Cross-validation is used during the calibration process to ensure the generalization ability of the model, and the cross-validation error does not exceed 8%. Preferably, the parametric adjustment algorithm in step 5 is based on B-spline curve interpolation to correct the outline of the cut piece. The shape of the cut piece is changed by adjusting the coordinates of the control points of the B-spline curve. The accuracy of the control point coordinates is ±0.1mm. The area of ​​the cut piece in the high strain region is increased by 5%-10%, and the area of ​​the cut piece in the low strain region is reduced by 3%-5%. The angle between the seam and the main strain direction of human body deformation does not exceed 30°. The elastic modulus adjustment range corresponding to the slip is: 0.5-1MPa when the slip is 3-5mm, and 1-1.5MPa when the slip is 5-8mm. Preferably, the physical parameters of the clothing fabric in step 6 include elastic modulus, thickness, and coefficient of friction, wherein the elastic modulus ranges from 0.1 to 5 MPa, the thickness ranges from 0.1 to 2 mm, the coefficient of friction ranges from 0.1 to 0.6, the simulation time step of the mass spring model does not exceed 0.01 s, and the output accuracy of the fabric stress value is ±0.01 MPa.

[0006] Preferably, the preset optimization target in step 7 is that the stress does not exceed 1.5MPa and the bonding gap does not exceed 1.5mm, the number of iterations does not exceed 5, the criterion for judging the stress concentration area is that the stress is not less than 2MPa, and the criterion for judging the bonding gap exceeding the standard is that the bonding gap is not less than 2mm. Preferably, the neural network algorithm in step 8 is the BP neural network algorithm, and the correction model reduces the prediction error by no less than 10%. The flexible wearable sensors transmit data synchronously to the acquisition terminal through a wireless transmission module. The wireless transmission module supports Bluetooth 5.0 or WiFi 6, and the data transmission delay does not exceed 50ms.

[0007] This invention discloses a system for dynamic adaptive clothing pattern generation based on soft tissue deformation and multi-pose anthropometric measurements. The system includes a data acquisition module, a human body deformation modeling module, an automatic pattern optimization module, a fabric simulation module, a user interaction module, a feedback correction module, and a power supply module. The data acquisition module is bidirectionally connected to the human body deformation modeling module. The human body deformation modeling module is bidirectionally connected to both the automatic pattern optimization module and the feedback correction module. The automatic pattern optimization module is unidirectionally connected to the fabric simulation module. The fabric simulation module is unidirectionally connected to both the user interaction module and the feedback correction module. The user interaction module is unidirectionally connected to the automatic pattern optimization module. The power supply module is electrically connected to all other modules to provide power. The data acquisition module is used to collect and synchronously transmit 3D data of the human body in multiple poses, as well as skin strain, slippage, and displacement data. The human body deformation modeling module is used to process the collected data, build and maintain a human body deformation database, and establish a layered mechanical model of soft tissue, realizing data registration optimization, model parameter calibration, and rapid database query. The pattern automatic optimization module is used to calculate the surface deformation data of the human body in the target pose, generate a dynamic surface model, correct the outline of the cut pieces, and optimize the seam layout and elastic fabric parameters. The fabric simulation module is used to receive the physical parameters of the fabric, simulate the fitting state and deformation effect of the clothing in multiple poses, and output stress distribution and fitting gap data. The user interaction module is used for users to input body shape parameters, select poses, view virtual try-on effects, and export pattern-related files, supporting real-time rendering and multi-view viewing. The feedback correction module is used to receive actual wearing feedback data, calculate the error between simulation and actual data, correct the mechanical model parameters, and supplement the database.

[0008] Preferably, the data acquisition module includes a multi-view 3D scanning unit, a flexible sensing unit, and a data transmission unit. The multi-view 3D scanning unit includes 3-8 synchronously triggered structured light 3D scanners and 2-4 depth cameras with a field of view ≥120°. The flexible sensing unit includes 16-32 flexible resistive strain sensors and medical-grade silicone fixation patches. The data transmission unit includes a Bluetooth 5.0 and WiFi 6 dual-mode transmission module. The 3D scanners are arranged around the human body within a range of 1.5-3m. The depth cameras are deployed in front of, behind, and on both sides of the human body. The strain sensors are attached to the shoulders, elbows, waist, hips, knees, ankles, and abdomen of the human body via silicone patches. The human body deformation modeling module includes a data preprocessing subunit, a database management subunit, and a mechanical modeling subunit. The data preprocessing subunit incorporates iterative nearest point algorithm and Poisson reconstruction algorithm modules. The database management subunit uses an encrypted MySQL database. The mechanical modeling subunit incorporates a finite element analysis module and calculation modules for Hill muscle model, Ogden hyperelastic model, and linear elastic model. The human body deformation modeling module is used to register, remove outliers, and optimize the topology of the collected 3D data, generate a structured mesh model, store, retrieve, and update a multidimensional human body deformation database containing at least 10,000 samples, and construct a soft tissue layered coupled mechanical model based on the preprocessed data and complete parameter calibration. The automatic pattern optimization module includes a dynamic surface calculation subunit, a piece adjustment subunit, and a seam and elastic fabric optimization subunit. The dynamic surface calculation subunit has a built-in finite element discretization processing module and an equilibrium equation solving module. The piece adjustment subunit has a built-in B-spline curve interpolation algorithm module. The seam and elastic fabric optimization subunit has a built-in strain and slippage analysis module. The automatic pattern optimization module is used to receive the mechanical model, calculate the surface strain distribution, tensile amount, and slippage trajectory under the target human body posture, generate a dynamic human body surface model, correct the piece contour based on the dynamic surface model, and determine the optimal seam layout and elastic fabric parameters. The fabric simulation module includes a fabric parameter input subunit, a simulation calculation subunit, and a result output subunit. The simulation calculation subunit has a built-in mass spring model calculation module, and the result output subunit includes a color cloud map rendering module. The fabric simulation module is used to receive physical parameters such as the elastic modulus and thickness of the fabric, simulate the fit, stretching and wrinkle distribution of the garment under multiple postures, and output stress distribution and fit gap data in the form of a color cloud map. The user interaction module includes a touch screen to display a parameter input interface, a posture selection interface, a virtual try-on interface, and a result export interface, and the virtual try-on interface has a built-in real-time rendering module; the user interaction module is used for users to input body shape parameters, select daily high-frequency postures, view virtual try-on effects, and export pattern files and pattern reports. The feedback correction module includes a feedback data acquisition subunit, an error analysis subunit, and a model update subunit. The feedback data acquisition subunit includes a wireless receiving module, and the model update subunit has a built-in BP neural network algorithm module. The feedback correction module is used to receive actual wearing strain data from the built-in sensors of the clothing, calculate the error between the simulation data and the actual data, correct the soft tissue mechanical model parameters, and supplement them to the human body deformation database. The power module includes an AC power supply interface, a DC power supply interface, and a battery.

[0009] Compared with the prior art, the present invention has the following significant advantages: Excellent dynamic adaptability: By collecting three-dimensional data and soft tissue deformation data of various typical human postures, a multi-posture human deformation database is established. Combined with a layered mechanical model, the deformation of the human body surface under different postures is accurately predicted, so that the generated clothing pattern can dynamically match the human body movement state and effectively avoid problems such as tightness, wrinkles, and slippage.

[0010] High modeling accuracy: The mechanical model adopts a layered coupling of muscle, fat and skin layers, combined with finite element analysis and calibration of a large number of samples. The model predicts deformation with small error, achieving a physical-level accurate description of soft tissue deformation and providing reliable data support for pattern optimization.

[0011] High degree of automation and personalization: It can automatically complete the correction of the cut piece outline, optimization of the seam layout and configuration of elastic fabric based on the body shape parameters or 3D scanning model input by the user, without the need for repeated manual adjustments. At the same time, it can adapt to users with different body shapes and different daily posture habits, achieving precise personalized customization.

[0012] Strong iterative optimization capability: By comparing actual wear feedback data with simulation results, the BP neural network algorithm is used to continuously correct the mechanical model parameters and continuously expand the human body deformation database, so that the accuracy of subsequent pattern generation can be continuously improved and adapt to more diverse body shapes and posture scenarios. Significant practicality and promotional value: The methodology is clear and the system structure is stable. It can be applied to scenarios such as clothing customization companies, intelligent fitting equipment, and sportswear R&D, effectively reducing enterprise production costs and return rates, improving user experience, and has broad industry application prospects. Attached Figure Description

[0013] Figure 1 This is a flowchart of the method of the present invention.

[0014] Figure 2 This is a schematic diagram of the human body in multiple poses according to the present invention.

[0015] Figure 3 This is a schematic diagram of the layered structure for soft tissue deformation modeling in this invention.

[0016] Figure 4 This is a schematic diagram of the dynamic pattern generation and fabric simulation of the present invention.

[0017] Figure 5 This is a diagram of the architecture of each module of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Method / Solution of this Invention: This invention discloses a dynamic adaptive clothing pattern generation method based on soft tissue deformation and multi-pose anthropometrics, comprising the following steps: Step 1, Multi-pose Human Body Data Acquisition: This involves acquiring 3D point cloud or mesh data of the human body in six typical postures: standing, bending, sitting, walking, arm raising, and twisting, using a data acquisition system composed of 3-8 synchronously triggered 3D scanners, 2-4 depth cameras, and 16-32 flexible wearable sensors. Specifically, when bending, the forward tilt angle is 30°-90°; when sitting, the knee bend angle is 90°; when walking, the step frequency is 50-120 steps / minute; when raising the arm, the angle between the arm and torso is 90°-180°; and when twisting, the left-right torso twist angle is 30°-60°. The flexible wearable sensors are... Flexible resistive strain sensors are attached to key deformation areas of the human body, such as the shoulder, elbow, waist, hip, knee, ankle, and abdomen. The strain measurement range is 0-50%, the slip measurement accuracy is ±0.1mm, and the displacement data sampling frequency is not less than 100Hz. The sensors transmit data synchronously to the acquisition terminal through a wireless transmission module that supports Bluetooth 5.0 or WiFi 6, with a data transmission delay of no more than 50ms. During the acquisition process, the 3D scanner uses structured light scanning technology with a scanning speed of not less than 10 frames / second, and the depth camera has a field of view of not less than 120° to ensure a full-body scan without blind spots. Step 2, Data Preprocessing and Database Construction: The multi-pose 3D data collected in Step 1 is registered using the iterative nearest point algorithm. Taking the standing 3D data as the baseline, other pose data are spatially aligned through point cloud feature point matching and rigid transformation iterative optimization. The registration error is controlled within ±0.3mm. After registration, the distance error of each corresponding point is calculated and outliers exceeding 3 times the average registration error are removed. Topology optimization is performed using the Poisson reconstruction algorithm. First, the normal vector of the registered point cloud data is estimated and oriented. Implicit functions are constructed and solved to obtain continuous surfaces. Then, a uniformly distributed structured mesh model is generated through adaptive sampling. The mesh cell size is uniformly 1-3mm. The deformation vector field, principal strain direction, and strain distribution matrix of the human body surface under each pose are extracted. Combined with body shape parameters such as height, weight, chest circumference, waist circumference, hip circumference, body fat percentage, and muscle mass, a multi-dimensional human deformation database with no less than 10,000 samples is established. Each sample in the database is associated with a pose category, body shape parameter group, and corresponding soft tissue deformation parameter set. Step 3, Soft Tissue Layered Mechanical Modeling: A multi-layer coupled mechanical model based on finite element analysis is adopted to divide human soft tissue into muscle, fat, and skin layers. The muscle layer uses the Hill muscle model, describing the active deformation caused by muscle contraction by inputting muscle activation parameters in the range of 0-1, and defining deformation anisotropy by combining muscle fiber orientation parameters preset through anatomical data and corrected for individual body shape. The fat layer uses the Ogden hyperelastic model, describing the nonlinear elastic deformation characteristics of fat under compression and tension by setting strain energy density function parameters adjusted according to body fat percentage (the higher the body fat percentage, the larger the parameter value). The skin layer uses a linear elastic model, describing the linear tensile and shear deformation of the skin by defining elastic modulus in the range of 1-5 MPa and Poisson's ratio in the range of 0.3-0.45. The model parameters are calibrated using the least squares method with sample data from the database. The mechanical parameters of each layer are determined by iterative calculation using the actual deformation data of the samples as the target values. The deformation prediction error of the calibrated model does not exceed 5%. Cross-validation is used during the calibration process to ensure the generalization ability of the model, and the cross-validation error does not exceed 8%. Step 4, Dynamic Pattern Parameter Calculation: Receive user-inputted body shape parameters or uploaded 3D scan model. Combined with 1-3 user-selected daily high-frequency postures, call the soft tissue mechanical model established in Step 3. Divide the human body surface into quadrilateral shell elements with side lengths of 2-5mm using finite element discretization. Establish element equilibrium equations based on the principle of virtual work. Substitute the mechanical parameters of each layer corresponding to the body shape and the constraints of the target posture. For example, use the Newton-Raphson iterative method to solve the equilibrium equations (iteration convergence condition is residual less than 1e-6) to obtain the strain tensor of each calculation element. Obtain the surface strain distribution through strain tensor decomposition. Calculate the stretching amount using the Euclidean distance formula based on the position coordinate difference between each element in the target posture and the reference posture. Based on the displacement vector difference between adjacent elements and time series data, obtain the sliding trajectory through B-spline curve interpolation (trajectory output accuracy is ±0.2mm), generating a dynamic human body surface model. Step 5, Automatic Adjustment of Garment Pattern: Based on the dynamic surface model from Step 4, the corresponding initial pattern library is called according to the garment type (e.g., top, pants, dress, etc.). The outline of the cut pieces is automatically corrected using a parametric adjustment algorithm based on B-spline curve interpolation. The shape of the cut pieces is changed by adjusting the coordinates of the control points of the B-spline curve with an accuracy of ±0.1mm to ensure smooth and continuous edges. The area of ​​the cut pieces is increased by 5%-10% for high-strain areas with surface strain not less than 15%, and the area of ​​the cut pieces is reduced by 3%-5% for low-strain areas with surface strain not greater than 5%. The seam layout is optimized so that the angle between the seam and the main strain direction of human body deformation does not exceed 30°. Elastic fabric blocks are placed in areas with slippage not less than 3mm. The elastic modulus of the elastic fabric blocks is adaptively adjusted according to the slippage, with an elastic modulus of 0.5-1MPa when the slippage is 3-5mm and 1-1.5MPa when the slippage is 5-8mm. Step 6, Fabric Physical Simulation Verification: Obtain the physical parameters of the clothing fabric, such as the elastic modulus in the range of 0.1-5MPa, the thickness in the range of 0.1-2mm, and the coefficient of friction in the range of 0.1-0.6. Use a mass spring model to simulate the fabric. The mass of each point of the fabric in the model is not less than 0.01g. The damping coefficient of the spring is adjusted according to the fabric type (0.1-0.3N・s / m for cotton fabrics and 0.05-0.1N・s / m for elastic fabrics). During the simulation, the effects of gravity, friction on the human body surface, and tension inside the fabric are considered. The simulation time step does not exceed 0.01s. Simulate the fit, stretching, and wrinkle distribution of the clothing under the high-frequency posture selected by the user. The simulation results are displayed in the form of a color cloud map to show the stress distribution and fit gap. Output the fabric stress value of each part with an accuracy of ±0.01MPa. Step 7, Pattern Iteration Optimization: If there are stress concentration areas with stress not less than 2MPa or bonding gaps exceeding the standard of not less than 2mm in the simulation results, automatically return to Step 5 to adjust the cutting parameters, seam positions or elastic fabric configuration, with no more than 5 iterations, until the preset optimization target of stress not exceeding 1.5MPa and bonding gap not exceeding 1.5mm is met. Step 8, Feedback Correction and Model Update: The fabric strain data during actual wear is collected by the flexible strain sensor built into the garment and compared with the simulation results in Step 6. The error value is calculated, and the parameters of the soft tissue mechanical model are corrected using the BP neural network algorithm. The prediction error of the corrected model is reduced by no less than 10%. The user's body shape parameters, posture data, actual deformation data, and optimized pattern parameters are added to the multidimensional human body deformation database in Step 2 to achieve model self-learning and continuous optimization.

[0020] The system content / solution of this invention: This invention discloses a dynamic adaptive clothing pattern generation system based on soft tissue deformation and multi-pose human body measurement, comprising a data acquisition module, a human body deformation modeling module, an automatic pattern optimization module, a fabric simulation module, a user interaction module, a feedback correction module, and a power supply module. The data acquisition module consists of a multi-view 3D scanning unit, a flexible sensing unit, and a data transmission unit. The multi-view 3D scanning unit includes 3-8 synchronously triggered structured light 3D scanners and 2-4 depth cameras with a field of view ≥120°. The 3D scanners have a scanning accuracy ≤0.1mm and a scanning speed ≥10 frames / second. The flexible sensing unit includes 16-32 flexible resistive strain sensors and medical-grade silicone fixation patches. The strain sensors have a thickness ≤0.5mm, a strain measurement range of 0-50%, and a response time ≤10ms. The data transmission unit includes a Bluetooth 5.0 and WiFi 6 dual-mode transmission module. The 3D scanners are arranged around the human body within a range of 1.5-3m. The depth cameras are deployed in front of, behind, and on both sides of the human body. The strain sensors are attached to the shoulders, elbows, waist, hips, knees, ankles, and abdomen of the human body via silicone patches. The data acquisition module is used to acquire three-dimensional point cloud or mesh data of human body in multiple poses, as well as skin strain, slip, and displacement data, and to transmit the acquired data synchronously. The human body deformation modeling module includes a data preprocessing subunit, a database management subunit, and a mechanical modeling subunit. The data preprocessing subunit has built-in iterative nearest point algorithm and Poisson reconstruction algorithm modules. The database management subunit uses a MySQL encrypted database. The mechanical modeling subunit has built-in finite element analysis module and Hill muscle model, Ogden hyperelastic model, and linear elastic model calculation modules. The human deformity modeling module is used to register, remove outliers and optimize topology of the collected 3D data, generate a structured mesh model, store, retrieve and update a multidimensional human deformity database containing at least 10,000 samples, support fast query by body shape parameters and posture category, with a response time of ≤1s, and construct a soft tissue layered coupled mechanical model based on the preprocessed data and complete parameter calibration. The automatic pattern optimization module includes a dynamic surface calculation subunit, a piece adjustment subunit, and a seam and elastic fabric optimization subunit. The dynamic surface calculation subunit has a built-in finite element discretization processing module and an equilibrium equation solving module. The piece adjustment subunit has a built-in B-spline curve interpolation algorithm module. The seam and elastic fabric optimization subunit has a built-in strain and slippage analysis module. The automatic pattern optimization module is used to receive the mechanical model, calculate the surface strain distribution, tensile amount, and slippage trajectory under the target human body posture, generate a dynamic human body surface model, correct the piece contour based on the dynamic surface model, and determine the optimal seam layout and elastic fabric parameters. The fabric simulation module includes a fabric parameter input subunit, a simulation calculation subunit, and a result output subunit. The simulation calculation subunit has a built-in mass spring model calculation module, and the result output subunit includes a color cloud map rendering module. The fabric simulation module receives physical parameters such as the fabric's elastic modulus and thickness, simulates the garment's fit, stretching, and wrinkle distribution under various postures, and outputs stress distribution and fit gap data in the form of a color cloud map. The user interaction module includes a touch screen with a parameter input interface, a posture selection interface, a virtual try-on interface, and a result export interface. The virtual try-on interface has a built-in real-time rendering module with a frame rate ≥30fps, and the result export interface supports DXF and PDF file output. The user interaction module allows users to input body shape parameters, select frequently used postures, view virtual try-on effects, and export pattern files and pattern reports. The feedback correction module includes a feedback data acquisition subunit, an error analysis subunit, and a model update subunit. The feedback data acquisition subunit includes a wireless receiving module with a receiving distance of ≥10 meters. The model update subunit has a built-in BP neural network algorithm module. The feedback correction module is used to receive actual wearing strain data from the built-in sensors of the clothing, calculate the error between the simulation data and the actual data, correct the soft tissue mechanical model parameters, and supplement them to the human body deformation database. The power module includes an AC power interface, a DC power interface, and a battery. The AC power interface is compatible with 220V±10%, the DC power interface is compatible with 12V±0.5V, and the battery capacity is ≥10000mAh with a battery life of ≥8 hours. The data acquisition module is bidirectionally connected to the human body deformation modeling module through the data transmission unit. The human body deformation modeling module is bidirectionally connected to the pattern automatic optimization module and the feedback correction module. The pattern automatic optimization module is unidirectionally connected to the fabric simulation module. The fabric simulation module is unidirectionally connected to the user interaction module and the feedback correction module. The user interaction module is unidirectionally connected to the pattern automatic optimization module. The power module is electrically connected to all other modules through wires.

[0021] In a preferred embodiment, each flexible resistive strain sensor in the flexible sensing unit of the data acquisition module integrates a temperature compensation module. The sensor's slip measurement accuracy is ±0.1mm, and the displacement data sampling frequency is ≥100Hz. The database management subunit of the human body deformation modeling module supports parallel storage and retrieval of multi-user data, and can simultaneously process the storage of body shape parameters, posture data, and deformation parameters of 5-10 users. It also has user privacy data encryption protection functions. The parameter calibration module of the mechanical modeling subunit adopts the least squares method. Combined with cross-validation, the model predicts deformation error ≤5%; in the pattern automatic optimization module's cut piece adjustment subunit, the control point coordinate adjustment accuracy of the B-spline curve interpolation algorithm is ±0.1mm, enabling adjustments of 5%-10% of the cut piece area in high-strain regions and 3%-5% in low-strain regions; the seam and elastic fabric optimization subunit can automatically match the elastic modulus of the elastic fabric based on the slippage amount, with a slippage amount of 3-5mm corresponding to an elastic modulus of 0.5-1MPa, and a slippage amount of 5-8mm corresponding to an elastic modulus of 1-1.5MPa; the fabric In the simulation calculation subunit of the material simulation module, the mass of the mass spring model is ≥0.01g, the spring damping coefficient for cotton fabric is 0.1-0.3N・s / m, and the spring damping coefficient for elastic fabric is 0.05-0.1N・s / m. The simulation time step is ≤0.01s, and the output accuracy of the fabric stress value is ±0.01MPa. The user interaction module supports access from the web and iOS / Android mobile terminals. The mobile terminal is adapted to devices with a screen size ≥4.7 inches. The virtual try-on interface supports zooming and rotation operations, and different postures can be switched in real time to view the details of clothing fit. The feedback data acquisition subunit of the feedback correction module can simultaneously receive wearing feedback data from up to 5 users to realize multi-user parallel testing. The prediction error of the mechanical model after correction by the model update subunit is reduced by ≥10%. The communication latency between the data acquisition module and the human body deformation modeling module is ≤50ms. The latency of the human body deformation modeling module transmitting the dynamic surface model to the pattern automatic optimization module is ≤1s. The frame rate of the simulation results transmitted from the fabric simulation module to the user interaction module is ≥30fps.

[0022] Example 1 (The following description uses specific numerical values ​​to illustrate the scheme) System Deployment and Data Acquisition System Deployment: A dynamic adaptive clothing pattern generation system is built. In the data acquisition module, three structured light 3D scanners are arranged around the human body within a 2m radius, and two depth cameras are deployed directly in front of and behind the human body to form a full-view scanning coverage. Sixteen flexible resistive strain sensors are attached to key deformation areas such as the shoulders, elbows, waist, hips, knees, ankles, and abdomen of the subject through medical-grade silicone patches. The data transmission unit uses Bluetooth 5.0 transmission mode to ensure that the acquired data is synchronized to the subsequent processing modules in real time.

[0023] Multi-posture data acquisition: Select one subject (e.g., height 175cm, weight 70kg, chest circumference 95cm, waist circumference 80cm, hip circumference 92cm, body fat percentage 18%) and collect 3D point cloud data of six typical postures: standing, bending at 45°, sitting (knee bent at 90°), walking (step frequency 80 steps / minute), raising arms at 120°, and torso twisting at 45°; at the same time, use flexible sensors to collect strain values, slippage and displacement data of key parts, with the sampling frequency set to 100Hz and the acquisition time lasting 30 seconds per posture to ensure data integrity and validity.

[0024] Data Preprocessing and Database Construction: Data Preprocessing: Through the data preprocessing subunit of the human body deformation modeling module, the iterative nearest point algorithm is used to register the 3D point cloud data of six poses. Taking the standing pose data as the benchmark, spatial alignment is achieved through point cloud feature point matching and rigid transformation iterative optimization, and the registration error is controlled within ±0.2mm. The Poisson reconstruction algorithm is used for topology optimization. First, the normal vector of the registered point cloud data is estimated and oriented, the implicit function is constructed and solved to obtain the continuous surface, and then a structured mesh model with a mesh cell size of 2mm is generated through adaptive sampling, and outliers exceeding 3 times the average registration error are removed.

[0025] Database Construction: Extract the deformation vector field, principal strain direction, and strain distribution matrix of the human body surface under various postures. Combine this with the subject's body shape parameters such as height, weight, circumference, body fat percentage, and muscle mass, and store it as a complete sample in a multidimensional human body deformation database (this database initially contains 10,000 sample data of different body shapes and postures). The database can be stored in an encrypted manner and supports fast queries by body shape parameters and posture categories, with a query response time of ≤1 second.

[0026] Soft tissue layered mechanical modeling constructs a layered coupled mechanical model of muscle, fat, and skin layers through the mechanical modeling sub-unit of the human body deformation modeling module: The muscle layer adopts the Hill muscle model, describing the active deformation caused by muscle contraction by inputting a muscle activation parameter of 0.6, and defining deformation anisotropy by combining muscle fiber orientation parameters preset based on human anatomy data and corrected for individual body shape; The fat layer adopts the Ogden hyperelastic model, describing the nonlinear elastic deformation characteristics of fat under compression and tension by setting strain energy density function parameters adjusted according to body fat percentage (the higher the body fat percentage, the larger the parameter value); The skin layer adopts a linear elastic model, defining an elastic modulus of 3 MPa and a Poisson's ratio parameter of 0.35 to describe the linear tensile and shear deformation of the skin. The least squares method is used, with actual deformation data of similar body shape samples in the database as the target value, and the mechanical parameters of each layer are determined by iterative calculation. The deformation prediction error of the calibrated model is 4.2%. Cross-validation is used during the calibration process to ensure the model's generalization ability, and the cross-validation error is 6.8%.

[0027] Dynamic pattern generation and optimization of dynamic surface model calculation: The user inputs their own body shape parameters (consistent with the test subject) through the touch screen of the interactive module and selects three daily high-frequency postures: walking, arm raised at 120°, and sitting. The dynamic surface calculation subunit of the pattern automatic optimization module calls the calibrated mechanical model, and divides the human body surface into quadrilateral shell elements with a side length of 3mm through finite element discretization. Based on the principle of virtual work, the element equilibrium equation is established. Substituting the mechanical parameters of each layer corresponding to the body shape and the constraint conditions of the target posture, the equilibrium equation is solved using the Newton-Raphson iterative method (the iteration convergence condition is that the residual is less than 1e-6), and the strain tensor of each calculation element is obtained. The surface strain distribution is obtained through strain tensor decomposition. According to the position coordinate difference of each element in the target posture and the reference posture, the Euclidean distance formula is used to calculate the stretching amount. Based on the displacement vector difference of adjacent elements and time series data, the sliding trajectory is obtained by fitting the B-spline curve interpolation method (the trajectory output accuracy is ±0.2mm), and the human body dynamic surface model is generated.

[0028] Automatic Pattern Adjustment: Based on the garment type of the sportswear, the system calls the corresponding initial pattern library and automatically corrects the outline of the cut pieces using a parametric adjustment algorithm based on B-spline curve interpolation. By adjusting the coordinates of the B-spline curve control points with an accuracy of ±0.1mm, the shape of the cut pieces is changed to ensure smooth and continuous edges. For high-strain areas such as the elbows and waist with surface strain ≥15%, the cut piece area is increased by 8%, while for low-strain areas such as the back with surface strain ≤5%, the cut piece area is reduced by 4%. The seam layout is optimized so that the angle between the seams at the shoulders and side seams and the principal strain direction of the human body deformation is ≤25°. Elastic fabric blocks are placed at the waist, cuffs, and other areas with slippage ≥3mm, where a slippage of 4mm at the waist corresponds to an elastic modulus of 0.8MPa, and a slippage of 6mm at the cuffs corresponds to an elastic modulus of 1.2MPa. V. Fabric Simulation and Iterative Optimization: Fabric Physical Simulation: Through the fabric parameter input subunit of the fabric simulation module, the physical parameters of the sports top fabric (elastic modulus 1.5MPa, thickness 0.8mm, coefficient of friction 0.3) are input. A mass spring model is used for fabric simulation. The mass of each point of the fabric in the model is set to 0.02g, and the damping coefficient of the spring is set to 0.15N・s / m. During the simulation, gravity, friction on the human body surface, and tension inside the fabric are considered. The simulation time step is set to 0.008s. The fit, stretching, and wrinkle distribution of the clothing under three high-frequency postures selected by the user are simulated. The simulation results are displayed in the form of a color cloud map to show the stress distribution and fit gap. The fabric stress value of each part is output with an accuracy of ±0.01MPa.

[0029] Iterative optimization: If the simulation results show stress concentration areas with stress ≥2MPa or bonding gaps ≥2mm exceeding the standard, the simulation automatically returns to the pattern adjustment step to adjust the cut piece parameters, seam positions, or elastic fabric configuration. In this embodiment, the first simulation shows that the elbow area stress is preferably 2.1MPa and the bonding gap is 1.8mm, which exceeds the preset standard. Therefore, the elbow cut piece area is increased by 3%, the elastic fabric coverage is adjusted, and the simulation is repeated. After the second simulation, the elbow stress drops to 1.3MPa and the bonding gap is 1.2mm, which meets the preset optimization target of stress ≤1.5MPa and bonding gap ≤1.5mm. The iteration terminates, and the number of iterations is 2 (not exceeding the upper limit of 5).

[0030] The feedback correction and model update process involves fabricating the optimized pattern into a physical sportswear top. Flexible strain sensors are embedded in key areas such as the elbows and waist of the garment. The user wears the top and performs actions such as walking, raising arms, and sitting. The feedback data acquisition subunit of the feedback correction module receives the actual wearing strain data from the built-in sensors and compares it with the simulation results output by the fabric simulation module to calculate the error value. The model update subunit uses a BP neural network algorithm to correct the parameters of the soft tissue mechanical model. The corrected model prediction error is reduced to 7.2%, a reduction of more than 10% compared to the original. At the same time, the user's body shape parameters, posture data, actual deformation data, and optimized pattern parameters are added to the multidimensional human body deformation database to achieve model self-learning and continuous optimization.

[0031] Through the above implementation process, the final sports top pattern can maintain good fit and comfort in various daily postures, effectively solving the problem of poor dynamic adaptability of traditional patterns, and verifying the practicality and reliability of the method and system of this invention.

[0032] Finally, it should be noted that the present invention is not limited to the above embodiments, and many variations are possible. All variations that can be directly derived or conceived by those skilled in the art from the disclosure of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A dynamic adaptive clothing pattern generation method based on soft tissue deformation and multi-pose anthropometrics, characterized in that, Includes the following steps: Step 1, Multi-pose human body data acquisition: Collect 3D point cloud or mesh data of the human body in various typical poses, and at the same time set up flexible wearable sensors at key deformation parts of the human body to collect strain values, slippage and displacement data of the skin surface. Step 2, Data Preprocessing and Database Construction: The multi-pose 3D data collected in Step 1 are registered, and a structured mesh model is generated through topology optimization. The deformation vector field, principal strain direction and strain distribution matrix of the human body surface under each pose are extracted. Combined with human body shape parameters, a multi-dimensional human body deformation database is established. Each sample in the database is associated with a pose category, body shape parameter group and corresponding soft tissue deformation parameter set. Step 3, Soft tissue layered mechanical modeling: Using a multi-layer coupled mechanical model based on finite element analysis, the human soft tissue is divided into muscle layer, fat layer and skin layer. The deformation characteristics of each layer are described by the corresponding mechanical model, and the model parameters are calibrated by sample data in the database. Step 4, Dynamic Pattern Parameter Calculation: Receive the body shape parameters input by the user or the uploaded 3D scan model, combine them with the user's selected daily high-frequency postures, call the soft tissue mechanical model established in Step 3, calculate the surface strain distribution, tensile amount and slip trajectory of each part of the human body under the target posture, and generate a dynamic surface model of the human body. Step 5, Automatic Adjustment of Garment Pattern: Based on the dynamic surface model in Step 4, the corresponding initial pattern library is called according to the garment type. The outline of the cut pieces is automatically corrected through a parametric adjustment algorithm, the seam layout is optimized, and elastic fabric blocks are configured in the corresponding parts. The elastic modulus of the elastic fabric blocks is adaptively adjusted according to the amount of slippage. Step 6, Fabric physical simulation verification: Obtain the physical parameters of the clothing fabric, use a mass spring model to simulate the fabric, simulate the fit, stretching and wrinkle distribution of the clothing under the high-frequency posture selected by the user, and output the fabric stress value of each part of the simulation results. Step 7, Pattern Iteration Optimization: If there are stress concentration areas or areas with excessive bonding gaps in the simulation results, automatically return to Step 5 to adjust the cut piece parameters, seam positions, or elastic fabric configuration until the preset optimization target is met; Step 8, Feedback Correction and Model Update: The fabric strain data during actual wear is collected by the flexible strain sensor built into the clothing, compared with the simulation results of Step 6, the error value is calculated, the parameters of the soft tissue mechanical model are corrected by a neural network algorithm, and the relevant user data is added to the human body deformation database of Step 2 to realize model self-learning.

2. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, The key deformation areas of the human body in Step 1 include the shoulder, elbow, waist, hip, knee, ankle and abdomen. The flexible wearable sensor is a flexible resistive strain sensor with a strain measurement range of 0-50%, a slip measurement accuracy of ±0.1mm, and a displacement data sampling frequency of not less than 100Hz. The typical postures in Step 1 include standing, bending over, sitting, walking, raising arms and twisting. When bending over, the forward tilt angle of the human body is 30°-90°. When sitting, the knee bend angle is 90°. When walking, the step frequency is 50-120 steps / minute. When raising arms, the angle between the arm and the torso is 90°-180°. When twisting, the left and right torso twists by 30°-60°.

3. A method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics as described in claim 1 or 2, characterized in that, In step 2, the registration uses the iterative nearest point algorithm, and the registration error is controlled within ±0.3mm. The topology optimization uses the Poisson reconstruction algorithm. The human body shape parameters include height, weight, chest circumference, waist circumference, hip circumference, body fat percentage, and muscle mass. The number of samples in the multidimensional human deformity database is no less than 10,000.

4. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, In step 3, the Hill muscle model is used to describe the contraction characteristics of the muscle layer, the Ogden hyperelastic model is used to describe the nonlinear deformation of the fat layer, and the linear elastic model is used to describe the tensile characteristics of the skin layer. The least squares method is used for model parameter calibration, with the actual deformation data of the samples in the database as the target value. The mechanical parameters of each layer are determined by iterative calculation. The deformation prediction error of the calibrated model does not exceed 5%. Cross-validation is used during the calibration process to ensure the generalization ability of the model, and the cross-validation error does not exceed 8%.

5. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, The parametric adjustment algorithm in step 5 is based on B-spline curve interpolation to correct the outline of the cut piece. The shape of the cut piece is changed by adjusting the coordinates of the control points of the B-spline curve. The accuracy of the control point coordinates is ±0.1mm. The area of ​​the cut piece in the high strain region is increased by 5%-10%, and the area of ​​the cut piece in the low strain region is reduced by 3%-5%. The angle between the seam and the main strain direction of human body deformation does not exceed 30°. The elastic modulus adjustment range corresponding to the slip is: 0.5-1MPa when the slip is 3-5mm, and 1-1.5MPa when the slip is 5-8mm.

6. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, The physical parameters of the clothing fabric in step 6 include elastic modulus, thickness, and coefficient of friction. The elastic modulus ranges from 0.1 to 5 MPa, the thickness ranges from 0.1 to 2 mm, the coefficient of friction ranges from 0.1 to 0.6, the simulation time step of the mass spring model does not exceed 0.01 s, and the output accuracy of the fabric stress value is ±0.01 MPa.

7. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, The preset optimization targets in step 7 are that the stress does not exceed 1.5MPa and the bonding gap does not exceed 1.5mm, the number of iterations does not exceed 5, the criteria for judging the stress concentration area is that the stress is not less than 2MPa, and the criteria for judging the bonding gap exceeding the standard is that the bonding gap is not less than 2mm.

8. The method for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics according to claim 1, characterized in that, The neural network algorithm in step 8 is the BP neural network algorithm. The correction model reduces the prediction error by no less than 10%. The flexible wearable sensors transmit data synchronously to the acquisition terminal through a wireless transmission module. The wireless transmission module supports Bluetooth 5.0 or WiFi 6, and the data transmission delay does not exceed 50ms.

9. A system for generating dynamic adaptive clothing patterns based on soft tissue deformation and multi-pose anthropometrics as described in any one of claims 1-8, characterized in that, It includes a data acquisition module, a human body deformation modeling module, a pattern automatic optimization module, a fabric simulation module, a user interaction module, a feedback correction module, and a power supply module. The data acquisition module is bidirectionally connected to the human body deformation modeling module, the human body deformation modeling module is bidirectionally connected to both the pattern automatic optimization module and the feedback correction module, the pattern automatic optimization module is unidirectionally connected to the fabric simulation module, the fabric simulation module is unidirectionally connected to both the user interaction module and the feedback correction module, the user interaction module is unidirectionally connected to the pattern automatic optimization module, and the power supply module is electrically connected to all other modules to provide power. The data acquisition module is used to collect and synchronously transmit 3D data of the human body in multiple poses, as well as skin strain, slippage, and displacement data. The human body deformation modeling module is used to process the collected data, build and maintain a human body deformation database, and establish a layered mechanical model of soft tissue, realizing data registration optimization, model parameter calibration, and rapid database query. The pattern automatic optimization module is used to calculate the surface deformation data of the human body in the target pose, generate a dynamic surface model, correct the outline of the cut pieces, and optimize the seam layout and elastic fabric parameters. The fabric simulation module is used to receive the physical parameters of the fabric, simulate the fitting state and deformation effect of the clothing in multiple poses, and output stress distribution and fitting gap data. The user interaction module is used for users to input body shape parameters, select poses, view virtual try-on effects, and export pattern-related files, supporting real-time rendering and multi-view viewing. The feedback correction module is used to receive actual wearing feedback data, calculate the error between simulation and actual data, correct the mechanical model parameters, and supplement the database.

10. A dynamic adaptive clothing pattern generation system based on soft tissue deformation and multi-pose anthropometrics as described in claim 9, characterized in that, The data acquisition module includes a multi-view 3D scanning unit, a flexible sensing unit, and a data transmission unit. The multi-view 3D scanning unit comprises 3-8 synchronously triggered structured light 3D scanners and 2-4 depth cameras with a field of view ≥120°. The flexible sensing unit comprises 16-32 flexible resistive strain sensors and medical-grade silicone fixation patches. The data transmission unit includes a Bluetooth 5.0 and WiFi 6 dual-mode transmission module. The 3D scanners are arranged around the human body within a range of 1.5-3m. The depth cameras are deployed in front of, behind, and on both sides of the human body. The strain sensors are attached to the shoulders, elbows, waist, hips, knees, ankles, and abdomen of the human body via silicone patches. The human body deformation modeling module includes a data preprocessing subunit, a database management subunit, and a mechanical modeling subunit. The data preprocessing subunit has built-in iterative nearest point algorithm and Poisson reconstruction algorithm modules. The database management subunit uses a MySQL encrypted database. The mechanical modeling subunit has built-in finite element analysis module and Hill muscle model, Ogden hyperelastic model, and linear elastic model calculation modules. The human body deformation modeling module is used to register, remove outliers and optimize topology of the collected three-dimensional data, generate a structured mesh model, store, retrieve and update a multi-dimensional human body deformation database containing at least 10,000 samples, and construct a soft tissue layered coupled mechanical model based on the preprocessed data and complete parameter calibration. The automatic pattern optimization module includes a dynamic surface calculation subunit, a piece adjustment subunit, and a seam and elastic fabric optimization subunit. The dynamic surface calculation subunit has a built-in finite element discretization processing module and an equilibrium equation solving module. The piece adjustment subunit has a built-in B-spline curve interpolation algorithm module. The seam and elastic fabric optimization subunit has a built-in strain and slippage analysis module. The automatic pattern optimization module is used to receive the mechanical model, calculate the surface strain distribution, tensile amount, and slippage trajectory under the target human body posture, generate a dynamic human body surface model, correct the piece contour based on the dynamic surface model, and determine the optimal seam layout and elastic fabric parameters. The fabric simulation module includes a fabric parameter input subunit, a simulation calculation subunit, and a result output subunit. The simulation calculation subunit has a built-in mass spring model calculation module, and the result output subunit includes a color cloud map rendering module. The fabric simulation module is used to receive physical parameters such as the elastic modulus and thickness of the fabric, simulate the fit, stretching and wrinkle distribution of the garment in multiple postures, and output stress distribution and fit gap data in the form of color cloud map. The user interaction module includes a touch screen to display a parameter input interface, a posture selection interface, a virtual try-on interface, and a result export interface, and the virtual try-on interface has a built-in real-time rendering module. The user interaction module is used to allow users to input body shape parameters, select daily high-frequency postures, view virtual try-on effects, and export pattern files and pattern reports. The feedback correction module includes a feedback data acquisition subunit, an error analysis subunit, and a model update subunit. The feedback data acquisition subunit includes a wireless receiving module, and the model update subunit has a built-in BP neural network algorithm module. The feedback correction module is used to receive actual wearing strain data from the built-in sensors of the clothing, calculate the error between the simulation data and the actual data, correct the soft tissue mechanical model parameters, and supplement them to the human body deformation database. The power module includes an AC power supply interface, a DC power supply interface, and a battery.

Images