Method and system for modeling luggage and accessories thereof based on three-dimensional point cloud

Abstract

The invention relates to the field of data processing, in particular to a three-dimensional point cloud-based luggage and accessory modeling method and system, and the method comprises the steps: obtaining and preprocessing the currently collected point cloud data of the luggage and accessories thereof; calculating a texture complexity coefficient of each point cloud data; performing density division on the point cloud data according to the texture complexity coefficient, dividing the point cloud data into a plurality of density categories, and calculating an optimal density value corresponding to each density category; and sampling the point cloud data according to the optimal density value corresponding to each density category to obtain a luggage and accessory model thereof. According to the method, sampling is performed according to the optimal density value of each density category, so that redundant data can be reduced while the detail features of the model are maintained, and the finally obtained luggage and accessory model thereof is more accurate.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to a modeling method and system for bags and their accessories based on three-dimensional point clouds. Background Art

[0002] In the design and production process of bags and accessories, there are extremely high requirements for the accuracy of 3D models. Among them, the modeling method based on 3D point cloud can provide a higher-precision 3D model, more accurately grasp the product's size, shape, texture and other information, thereby improving product design quality and production efficiency.

[0003] Prior art, such as the patent application document with publication number CN117057206A, discloses an intelligent modeling method and system for three-dimensional molds, which includes: obtaining a calibrated mold image set through a preset camera, and partitioning the calibrated mold image set to extract feature point clouds, thereby obtaining mold point cloud data; performing three-dimensional point cloud densification on the mold point cloud data, thereby obtaining three-dimensional point cloud data; constructing a mold three-dimensional model based on the three-dimensional point cloud data, and iteratively optimizing the parameters of the mold three-dimensional model to obtain a best-fit three-dimensional model; parameterizing the best-fit three-dimensional model to obtain a parameterized model; obtaining mold demand data, and optimizing the parameterized model as a first mold model according to the mold demand data to obtain an optimized mold model; performing mechanical simulation performance simulation on the optimized mold model to obtain mold working simulation data; performing a second mold model optimization on the optimized mold model according to the mold working simulation data to obtain an optimal mold model, and converting the manufacturing parameters of the optimal mold model to obtain mold manufacturing parameter data.

[0004] However, in the above intelligent modeling method, for complex molds, their features are irregular, etc., which causes the densified point cloud data to contain a large amount of redundant information, increasing the difficulty and amount of calculation for subsequent processing, thereby reducing the construction accuracy and efficiency of the three-dimensional model. Summary of the invention

[0005] In order to solve the technical problem that the densified point cloud data contains redundant information, resulting in reduced construction accuracy and efficiency of three-dimensional modeling, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a modeling method for bags and their accessories based on three-dimensional point clouds includes: Acquire and preprocess the point cloud data of the currently collected luggage and its accessories; Calculate the texture complexity coefficient of each point cloud data; Density division of point cloud data is performed according to the texture complexity coefficient, the point cloud data is divided into multiple density categories, and the optimal density value corresponding to each density category is calculated; The point cloud data is sampled according to the optimal density value corresponding to each density category to obtain the model of the luggage and its accessories.

[0007] The present invention first calculates the texture complexity coefficient of the point cloud data, which can reasonably determine the boundary and range of the density category, and then samples different densities in areas of different complexity in a targeted manner, avoiding unnecessary waste of computing resources. Further sampling is performed according to the optimal density value of each density category, which can ensure that while maintaining the detailed features of the model, redundant data is reduced, making the final model of luggage and its accessories more accurate.

[0008] Preferably, the process of obtaining the texture complexity coefficient of the point cloud data includes: Take any point cloud data as the target point cloud, obtain the neighborhood point cloud of the target point cloud, and obtain the normal vectors of the target point cloud and the neighborhood point cloud; calculate the mean of the variance of the normal vectors of the neighborhood point cloud in three dimensions, and negatively correlate the mean of the variance of the normal vectors of the neighborhood point cloud in three dimensions to obtain the consistency of the normal vector of the target point cloud; cluster all point cloud data, count the total number of point cloud data in the category where the target point cloud is located, and calculate the point cloud ratio within the category of the target point cloud; wherein each category in the clustering result corresponds to a normal plane; The product of the normal vector consistency of the target point cloud and the ratio of the point cloud within the category is used as the texture complexity coefficient.

[0009] By calculating the mean of the variance of the normal vectors of the neighborhood point cloud in three dimensions and performing negative correlation mapping, the normal vector consistency of the target point cloud can be obtained. This indicator reflects the smoothness of the local surface of the point cloud and the consistency of the geometric features. The higher the normal vector consistency, the smoother the local surface of the point cloud and the more consistent the geometric features, which helps to more accurately evaluate the complexity of the texture. By clustering all point cloud data and counting the total number of point cloud data in the category where the target point cloud is located, the point cloud ratio within the category of the target point cloud can be calculated. This ratio reflects the similarity and degree of aggregation between the target point cloud and its surrounding point clouds. The higher the point cloud ratio within the category, the more similar the texture features of the target point cloud and its surrounding point clouds are, which helps to further refine the evaluation of texture complexity. By combining normal vector consistency and point cloud ratio within a category, a rapid assessment of the texture complexity of point cloud data can be achieved. This process avoids the complex calculation and feature extraction steps in traditional texture analysis methods, thereby improving processing efficiency.

[0010] Preferably, the density division of the point cloud data according to the texture complexity coefficient to divide the point cloud data into a plurality of density categories comprises: Point cloud data whose texture complexity coefficient is less than a preset first threshold is classified into a low-density category; point cloud data whose texture complexity coefficient is greater than or equal to the preset first threshold and less than a preset second threshold is classified into a medium-density category; point cloud data whose texture complexity coefficient is greater than or equal to the preset second threshold is classified into a high-density category.

[0011] The point cloud data is density-divided according to the texture complexity coefficient and divided into low-density category, medium-density category and high-density category, which has the functions of optimizing data processing strategy, improving model construction quality and supporting multi-scale analysis and visualization.

[0012] Preferably, calculating the optimal density value corresponding to each density category includes: For each density category, multiple sampling intervals are set, and the point cloud data is evenly sampled at each interval; the average value of the error between the corresponding point cloud fitting surface after all interval sampling and the original point cloud fitting surface is calculated; A negative correlation mapping is performed on the average value of the error between the corresponding point cloud fitting surface and the original point cloud fitting surface after all interval sampling, and the interval corresponding to the minimum value after the negative correlation mapping is obtained, and the interval is used as the optimal density value of the corresponding density category.

[0013] By setting multiple sampling intervals for each density category and performing even sampling, the characteristics of the point cloud data under different sampling densities can be analyzed in more detail. Then, the average value of the error between the corresponding point cloud fitting surface after all interval sampling and the original point cloud fitting surface can be calculated. The impact of sampling density on the fitting accuracy of point cloud data can be objectively evaluated. By mapping the negative correlation and finding the interval corresponding to the minimum value, the optimal density value can be directly found, avoiding exhaustive search under all possible sampling densities.

[0014] Preferably, the interval corresponding to the minimum value after obtaining the negative correlation mapping is expressed by a relational expression: ; In the formula, For the The optimal density value for density categories, For the The average error between the surface fitted by the sampling points corresponding to the interval z and the surface fitted by the original point cloud data within the density category, is an exponential function with the natural constant e as base; represents a function, that is, inputting the minimum weighted error and outputting the corresponding optimal density value; z represents the sampling interval.

[0015] Preferably, sampling the point cloud data according to the optimal density value corresponding to each density category to obtain the model of the luggage and its accessories includes: Select voxel grid sampling and set the voxel size of the density category to the resolution corresponding to the optimal density value of the density category, and downsample the currently collected point cloud data of the luggage and its accessories; The sampled point cloud data is denoised and filtered, and the processed point cloud data is modeled using 3D modeling software.

[0016] By selecting the voxel grid sampling method and setting the corresponding voxel size (i.e., resolution) according to the optimal density value of the density category, accurate sampling of point cloud data can be achieved. This sampling strategy can ensure that the amount of data is effectively reduced and the processing speed is improved while maintaining the characteristics of the point cloud data. At the same time, since the sampling process is based on the optimal density value, the detailed information of the point cloud data can be retained to the maximum extent, thereby improving the accuracy and realism of the model.

[0017] Preferably, the process of obtaining the texture complexity coefficient of the point cloud data includes: Select any point cloud data as the target point cloud, construct the neighborhood of the target point cloud according to the preset number of neighborhood points, and calculate the average angle between the normal vectors of all neighborhood points and the normal vector of the target point cloud; Calculate the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud; The product of the angle mean and the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud is normalized to obtain the texture complexity coefficient of the target point cloud.

[0018] This operation not only considers the mean angle between the target point cloud and its neighboring point cloud normal vectors, but also introduces the standard deviation of the curvature of the point cloud data in the neighborhood. The mean angle reflects the geometric consistency of the local surface of the target point cloud, while the standard deviation of the curvature reflects the degree of geometric shape change of the point cloud data in the neighborhood. After the product of the two is normalized, the texture complexity of the point cloud can be more comprehensively evaluated.

[0019] In the second aspect, a luggage and accessories modeling system based on three-dimensional point clouds includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned luggage and accessories modeling method based on three-dimensional point clouds is implemented.

[0020] The beneficial effects of the present invention are: The present invention can more accurately capture the detailed features of luggage and its accessories by performing texture complexity analysis on point cloud data, divide the density according to the texture complexity coefficient, divide the point cloud data into different density categories, perform targeted sampling, optimize resource allocation, and improve modeling efficiency and accuracy; select the optimal sampling density by minimizing the error to ensure that the sampled point cloud data reaches the best balance between fitting accuracy and computational efficiency, perform sampling and modeling according to the optimal density value, generate high-quality point cloud data, thereby quickly constructing a three-dimensional model, while retaining the details of complex areas and improving the overall modeling quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 It is a method flow chart of steps S1 to S4 in the method for modeling bags and their accessories based on three-dimensional point clouds in an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The embodiment of the present invention discloses a modeling method for bags and their accessories based on three-dimensional point clouds, referring to Figure 1 , including steps S1 to S4, which are specifically as follows: S1: Acquire and preprocess the point cloud data of the currently collected luggage and its accessories.

[0024] Point cloud data can accurately reflect the three-dimensional shape and surface features of luggage and its accessories. By collecting point cloud data, operators can quickly obtain a three-dimensional model of the product for further design optimization and improvement.

[0025] Specifically, point cloud acquisition equipment (such as binocular cameras, point cloud scanners, etc.) is used to collect data on luggage and its accessories.

[0026] The point cloud data collected by the acquisition device is a set of three-dimensional points, each of which contains its coordinates (x, y, z) in space, which are used to describe the surface shape of luggage and its accessories.

[0027] After collecting the point cloud data, it needs to be preprocessed, including denoising and outlier filtering, to reduce errors and interference in the data and make the point cloud data more accurately reflect the true surface features of luggage and its accessories.

[0028] S2: Calculate the texture complexity coefficient of each point cloud data.

[0029] After collecting the three-dimensional point cloud data of luggage and its accessories, high-density point cloud data is usually required to obtain a high-precision model. However, high-density point cloud data is less efficient when modeling. Therefore, it is necessary to sample the dense point cloud to reduce the point cloud density and thus improve the modeling efficiency.

[0030] In addition, the surface of bags has both flat areas and complex concave and convex patterns. If too dense a point cloud is used, data redundancy will result; if too dense a point cloud is used, the complexity of the concave and convex patterns will be reduced. Therefore, by calculating the texture complexity coefficient, the point cloud density can be dynamically adjusted to optimize the model.

[0031] Specifically, first, the normal vector of each point cloud data (i.e., the vector perpendicular to the plane where the point cloud data is located) is calculated to preliminarily understand the local surface features of each point cloud data.

[0032] Then, find the point cloud data in the neighborhood corresponding to each point cloud data, calculate the variance of the normal vectors of these neighborhood point cloud data in three dimensions (x, y, z), and take the average of the variances of these three dimensions to get the mean variance of the normal vectors of the neighborhood of each point cloud data. The smaller the mean variance, the more concentrated the distribution of the corresponding point cloud data in each dimension, that is, the higher the consistency. Conversely, the larger the mean variance, the lower the consistency. Therefore, the mean variance of the normal vectors of the neighborhood of each point cloud data is negatively correlated to obtain the normal vector consistency of the corresponding point cloud data.

[0033] For example, take any point cloud data as an example and use it as the target point cloud, obtain the neighborhood data of the target point cloud in three dimensions (a total of 26 neighborhood data), calculate the variance of all normal vectors in each dimension, and take the mean of the variance of the normal vectors in the three dimensions as a negative correlation map to obtain the normal vector consistency of the target point cloud, that is, satisfy the relationship:

[0034] In the formula, For the The consistency of normal vectors of point cloud data. For the The variance mean of the normal vectors of all neighborhood data corresponding to a point cloud data.

[0035] Secondly, a clustering method (such as mean shift or DBSCAN) is used to classify each point cloud data and its normal vector, and point cloud data with similar normal vectors are classified into the same category. These point cloud data can be approximately considered to be on the same plane.

[0036] Further confirming the above Which category does the point cloud data belong to (each category represents a normal plane), and count the The total number of point cloud data in the category of the point cloud data is calculated. The ratio of the total number of point cloud data in the category of the point cloud data to all point cloud data, and this ratio is used as the The point cloud ratio within the category of point cloud data satisfies the relationship:

[0037] In the formula, For the The point cloud ratio within the category of point cloud data, For the The total number of point cloud data in the category where the point cloud data belongs, is the total number of point cloud data collected by S1 above.

[0038] in, The larger the The more point cloud data are in the same normal plane, the higher the point cloud density on the plane is, so the required sampling density can be reduced; conversely, The smaller the The fewer the number of point cloud data points on the same normal plane, the lower the point cloud density on that plane, so a higher sampling density may be required to ensure the integrity and accuracy of the data.

[0039] Finally, the The product of the normal vector consistency of the point cloud data and the point cloud ratio within the category is taken as the first The texture complexity coefficient of the point cloud data satisfies the relationship:

[0040] In the formula, For the The texture complexity coefficient of the point cloud data is For the The consistency of normal vectors of point cloud data. For the The point cloud ratio within the category of the point cloud data.

[0041] No. The greater the texture complexity coefficient of the point cloud data, the The more complex the surface features of the area where the point cloud data is located, the higher the point cloud density is required.

[0042] Furthermore, repeat the above steps for all point cloud data to obtain value.

[0043] In another embodiment, another method for calculating the texture complexity coefficient of point cloud data is provided, and the specific steps are as follows: Select any point cloud data as the target point cloud, build the neighborhood of the target point cloud according to the preset number of neighborhood points (set according to the actual situation, such as setting a smaller number of neighborhood points in areas with higher density; and setting a larger number of neighborhood points in areas with lower density or more noise to include enough neighborhood points), and calculate the average angle between the normal vectors of all neighborhood points and the normal vector of the target point cloud; Calculate the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud; The product of the angle mean and the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud is normalized to obtain the texture complexity coefficient of the target point cloud.

[0044] S3: Density division of point cloud data is performed according to the texture complexity coefficient, the point cloud data is divided into multiple density categories, and the optimal density value corresponding to each density category is calculated.

[0045] The texture complexity coefficient of the point cloud data calculated according to the above S2 can reflect the complexity difference of the point cloud data in the overall texture, such as whether it is dense or sparse, but it cannot be accurate to the specific density value.

[0046] First, the point cloud data is preliminarily divided into multiple density categories (such as dense, medium, sparse, etc.) according to the texture complexity coefficient of each point cloud data. Exemplarily, the point cloud data with a texture complexity coefficient less than a preset first threshold is divided into a low density category; the point cloud data with a texture complexity coefficient greater than or equal to the preset first threshold and less than a preset second threshold is divided into a medium density category; the point cloud data with a texture complexity coefficient greater than or equal to the preset second threshold is divided into a high density category. The first threshold is 0.3 and the second threshold is 0.7.

[0047] Furthermore, a surface fitting is performed for each density category, and the surface fitting error under different sampling intervals is calculated. The specific steps are as follows: Set the sampling density of each category of point cloud data in space, such as taking a point every z units. For example, for the density categories, try multiple different z values, such as z = 1, 2, 3, ..., 10. Further, use even sampling (i.e., equal interval sampling) to calculate the average error between the surface fitted by the sampling points corresponding to the interval z and the surface fitted by the original point cloud data, and mark it as The calculated error is used to determine the effect of the current sampling interval. It should be noted that the calculation of the fitting error is a prior art and will not be described in detail here.

[0048] The above is under interval z sampling The larger it is, the greater the error after sampling, that is, the worse the uniform sampling effect of the current interval z is.

[0049] In order to realize a sparse and high-precision bag model, it is necessary to find a sampling interval z such that, Relatively small.

[0050] Specifically, Perform negative correlation mapping (i.e. multiply by ), that is, when When it is larger, due to The attenuation effect of The value of will decrease; when z is larger (that is, the sampling interval is larger), It will also decrease, which helps to find a balance between sparse sampling and accuracy.

[0051] Furthermore, by obtaining The z value corresponding to the minimum value of is taken as the optimal density value of this category. The z value corresponding to the minimum value of satisfies the relationship:

[0052] In the formula, For the The optimal density value for density categories, For the The average error between the surface fitted by the sampling points corresponding to the interval z and the surface fitted by the original point cloud data within the density category, is an exponential function with the natural constant e as base, Represents a function that accepts an input (i.e., the minimum weighted error) and outputs a result (i.e., the optimal density value ).

[0053] S4: Sample the point cloud data according to the optimal density value corresponding to each density category to obtain the model of the luggage and its accessories.

[0054] According to the characteristics of the point cloud data and the modeling requirements, the optimal density value of each density category is determined through the operations of the above steps S1 to S3.

[0055] An embodiment of the present invention also discloses a luggage and accessories modeling system based on three-dimensional point clouds, including a processor and a memory, wherein the memory stores computer program instructions. When the computer program instructions are executed by the processor, a luggage and accessories modeling method based on three-dimensional point clouds according to the present invention is implemented.

[0056] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0057] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module or both. Any such computer storage medium may be part of a device or accessible or connectable to a device.

[0058] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0059] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A modeling method for bags and their accessories based on three-dimensional point clouds, characterized in that: include: Acquire and preprocess the point cloud data of the currently collected luggage and its accessories; Calculate the texture complexity coefficient of each point cloud data; Density division of point cloud data is performed according to the texture complexity coefficient, the point cloud data is divided into multiple density categories, and the optimal density value corresponding to each density category is calculated; The point cloud data is sampled according to the optimal density value corresponding to each density category to obtain the model of the luggage and its accessories.

2. The modeling method of bags and their accessories based on three-dimensional point cloud according to claim 1 is characterized in that: The process of obtaining the texture complexity coefficient of the point cloud data includes: Take any point cloud data as the target point cloud, obtain the neighborhood point cloud of the target point cloud, and obtain the normal vectors of the target point cloud and the neighborhood point cloud; calculate the mean of the variance of the normal vectors of the neighborhood point cloud in three dimensions, and negatively correlate the mean of the variance of the normal vectors of the neighborhood point cloud in three dimensions to obtain the consistency of the normal vector of the target point cloud; cluster all point cloud data, count the total number of point cloud data in the category where the target point cloud is located, and calculate the point cloud ratio within the category of the target point cloud; wherein each category in the clustering result corresponds to a normal plane; The product of the normal vector consistency of the target point cloud and the ratio of the point cloud within the category is used as the texture complexity coefficient.

3. The method for modeling bags and their accessories based on three-dimensional point cloud according to claim 2, characterized in that: The density division of the point cloud data according to the texture complexity coefficient, and the division of the point cloud data into multiple density categories include: Point cloud data whose texture complexity coefficient is less than a preset first threshold is classified into a low-density category; point cloud data whose texture complexity coefficient is greater than or equal to the preset first threshold and less than a preset second threshold is classified into a medium-density category; point cloud data whose texture complexity coefficient is greater than or equal to the preset second threshold is classified into a high-density category.

4. The modeling method of bags and their accessories based on three-dimensional point cloud according to claim 3 is characterized in that: Calculating the optimal density value corresponding to each density category includes: For each density category, multiple sampling intervals are set, and the point cloud data is evenly sampled at each interval; the average value of the error between the corresponding point cloud fitting surface after all interval sampling and the original point cloud fitting surface is calculated; A negative correlation mapping is performed on the average value of the error between the corresponding point cloud fitting surface and the original point cloud fitting surface after all interval sampling, and the interval corresponding to the minimum value after the negative correlation mapping is obtained, and the interval is used as the optimal density value of the corresponding density category.

5. The modeling method of bags and their accessories based on three-dimensional point cloud according to claim 4 is characterized in that: The interval corresponding to the minimum value after obtaining the negative correlation mapping is expressed by the relational expression: ; In the formula, For the The optimal density value for density categories, For the The average error between the surface fitted by the sampling points corresponding to the interval z and the surface fitted by the original point cloud data within the density category, is an exponential function with the natural constant e as base; represents a function, that is, inputting the minimum weighted error and outputting the corresponding optimal density value; z represents the sampling interval.

6. The method for modeling bags and their accessories based on three-dimensional point cloud according to claim 4, characterized in that: The point cloud data is sampled according to the optimal density value corresponding to each density category to obtain the model of the luggage and its accessories, including: Select voxel grid sampling and set the voxel size of the density category to the resolution corresponding to the optimal density value of the density category, and downsample the currently collected point cloud data of the luggage and its accessories; The sampled point cloud data is denoised and filtered, and the processed point cloud data is modeled using 3D modeling software.

7. The modeling method of bags and their accessories based on three-dimensional point cloud according to claim 1, characterized in that: The process of obtaining the texture complexity coefficient of the point cloud data includes: Select any point cloud data as the target point cloud, construct the neighborhood of the target point cloud according to the preset number of neighborhood points, and calculate the average angle between the normal vectors of all neighborhood points and the normal vector of the target point cloud; Calculate the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud; The product of the angle mean and the standard deviation of the curvature of all point cloud data in the neighborhood of the target point cloud is normalized to obtain the texture complexity coefficient of the target point cloud.

8. The luggage and accessories modeling system based on three-dimensional point cloud is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for modeling bags and their accessories based on three-dimensional point clouds according to any one of claims 1 to 7 is implemented.

Images