Three-dimensional scene simulation method and system based on 3D Gaussian sputtering

Abstract

The invention provides a three-dimensional scene simulation method and system based on 3D Gaussian sputtering in the technical field of physical simulation. The method comprises the steps that S1, an initial point cloud model is obtained based on reconstruction of all object images; s2, based on the initial point cloud model, mask label reconstruction is carried out to obtain a triangular mesh model, and a tetrahedral mesh model is obtained through expansion; s3, creating a 3D Gaussian sputtering model based on the triangular mesh model through a 3D Gaussian kernel function; s4, through implicit integration and finite element modes, controlling the tetrahedral mesh model to simulate the deformation and motion effect of the simulation target based on physical attributes, and collecting the deformation gradient and vertex coordinates of each tetrahedral mesh in the deformation process; s5, synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient; and S6, the deformed 3D Gaussian sputtering model is mapped to the tetrahedral mesh model based on the vertex coordinates, and a simulation result is obtained through rendering. The method has the advantage that the three-dimensional scene simulation precision is greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of physical simulation technology, and in particular to a three-dimensional scene simulation method and system based on 3D Gaussian sputtering. Background Art

[0002] The emergence of 3D Gaussian Sputtering technology has successfully transformed the traditional 3D reconstruction method. 3D Gaussian Sputtering not only significantly improves the computational speed of reconstructing scenes from multi-view images, but also provides a novel 3D scene geometric expression.

[0003] However, some of the current scene simulations based on 3D Gaussian sputtering only focus on the changes of the scene over time, but ignore the physical laws behind the changes; some focus on physics-based motion and deformation, but their simulation results are still difficult to achieve high realism, and they are also unable to handle the collision detection and response problems between multiple reconstructed targets well. That is, the current scene simulation based on 3D Gaussian sputtering has poor simulation accuracy.

[0004] Therefore, how to provide a three-dimensional scene simulation method and system based on 3D Gaussian sputtering to improve the accuracy of three-dimensional scene simulation has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a three-dimensional scene simulation method and system based on 3D Gaussian sputtering, so as to improve the simulation accuracy of the three-dimensional scene.

[0006] In a first aspect, the present invention provides a three-dimensional scene simulation method based on 3D Gaussian sputtering, comprising the following steps:

[0007] Step S1, obtaining physical images of the simulation target at different viewing angles, marking the simulation target in each of the physical images to obtain a mask label, and reconstructing an initial point cloud model based on each of the physical images;

[0008] Step S2: reconstructing a triangular mesh model of the simulation target based on the initial point cloud model and the mask label, and expanding each triangular mesh model to obtain a tetrahedral mesh model;

[0009] Step S3, creating a corresponding 3D Gaussian sputtering model based on the triangular mesh model through a 3D Gaussian kernel function;

[0010] Step S4, configuring the physical properties of the simulation target, controlling the tetrahedral mesh model to simulate the deformation and motion effect of the simulation target based on the physical properties through implicit integration and finite element methods, and collecting the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model during the deformation process;

[0011] Step S5, synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient;

[0012] Step S6: Map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

[0013] Furthermore, the step S1 is specifically as follows:

[0014] Acquire real images of the simulated target at different viewing angles through a camera, and store each of the real images in real time to a preset image storage path; label the simulated target in each of the real images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path;

[0015] Image features are extracted from each of the physical object images through COLMAP, and each of the image features is input into Structure-from-Motion to reconstruct an initial point cloud model, and the initial point cloud model is stored in real time to a preset model storage path.

[0016] Furthermore, the step S4 is specifically as follows:

[0017] Configure the physical properties of the simulation target based on the JSON file, where the physical properties include at least material type, material parameters, and stress conditions;

[0018] Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

[0019] Furthermore, the step S5 is specifically as follows:

[0020] The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

[0021] Furthermore, the step S6 is specifically as follows:

[0022] The deformed 3D Gaussian sputtering model is mapped to a tetrahedral mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results.

[0023] In a second aspect, the present invention provides a three-dimensional scene simulation system based on 3D Gaussian sputtering, comprising the following modules:

[0024] An initial point cloud model creation module is used to obtain physical images of the simulation target from different perspectives, annotate the simulation target in each physical image to obtain a mask label, and reconstruct the initial point cloud model based on each physical image;

[0025] A tetrahedral mesh model creation module, used to reconstruct a triangular mesh model of a simulation target based on the initial point cloud model and the mask label, and expand each triangular mesh model to obtain a tetrahedral mesh model;

[0026] A 3D Gaussian sputtering model creation module, used to create a corresponding 3D Gaussian sputtering model based on the triangular mesh model by using a 3D Gaussian kernel function;

[0027] A tetrahedral mesh model deformation module is used to configure the physical properties of the simulation target, and control the deformation and motion effects of the simulation target based on the physical properties by implicit integration and finite element methods. During the deformation process, deformation gradients and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are collected;

[0028] A 3D Gaussian sputtering model deformation module, used for synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient;

[0029] The simulation rendering module is used to map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

[0030] Furthermore, the initial point cloud model creation module is specifically used for:

[0031] Acquire real images of the simulated target at different viewing angles through a camera, and store each of the real images in real time to a preset image storage path; label the simulated target in each of the real images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path;

[0032] Image features are extracted from each of the physical images through COLMAP, and each of the image features is input into Structure-from-Motion to reconstruct an initial point cloud model, and the initial point cloud model is stored in real time to a preset model storage path.

[0033] Furthermore, the tetrahedral mesh model deformation module is specifically used for:

[0034] Configure the physical properties of the simulation target based on the JSON file, where the physical properties include at least material type, material parameters, and stress conditions;

[0035] Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

[0036] Furthermore, the 3D Gaussian sputtering model deformation module is specifically used for:

[0037] The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

[0038] Furthermore, the simulation rendering module is specifically used for:

[0039] The deformed 3D Gaussian sputtering model is mapped to a tetrahedral mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results.

[0040] The advantages of the present invention are:

[0041] 1. By acquiring real images of the simulation target from different perspectives, the simulation target in each real image is marked to obtain a mask label, and the initial point cloud model is reconstructed based on each real image; then the triangular mesh model of the simulation target is reconstructed based on the initial point cloud model and the mask label, and each triangular mesh model is expanded to obtain a tetrahedral mesh model; then the corresponding 3D Gaussian sputtering model is created based on the triangular mesh model through the 3D Gaussian kernel function; then the physical properties of the simulation target are configured, and the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effect of the simulation target through implicit integration and finite element methods. During the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are collected; then the 3D Gaussian sputtering model is synchronously deformed based on the deformation gradient, and finally The deformed 3D Gaussian sputtering model is mapped to the tetrahedral mesh model based on the vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results; that is, by marking the simulation target, the focus is on the reconstruction of the simulation target and the interference of background redundant information is ignored; the tetrahedral mesh model is controlled through the configured physical properties, combined with implicit integration and finite element methods to simulate the deformation and motion effects of the simulation target, and the physical laws behind the changes of the simulation target are fully considered; collision detection and response are performed through penalty force or IPC, which can well handle the collision detection and response problems between multiple reconstructed targets (simulation results), and ultimately greatly improve the simulation accuracy of the three-dimensional scene.

[0042] 2. Through each time step in the deformation process, the energy equation is solved to obtain the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model, and the deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model to synchronously deform the 3D Gaussian sputtering model, thereby greatly improving the timeliness of the simulation.

[0043] 3. The tetrahedral mesh model is obtained by expanding the triangular mesh model. The tetrahedral mesh model is used as the main model of the simulation. It can be applied to implicit integration and finite element simulation methods to solve the physical motion and deformation equations as accurately as possible while supporting a large time step and ensuring system stability, thereby obtaining realistic simulation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The present invention will be further described below in conjunction with embodiments with reference to the accompanying drawings.

[0045] Figure 1 It is a flow chart of a three-dimensional scene simulation method based on 3D Gaussian sputtering of the present invention.

[0046] Figure 2It is a structural schematic diagram of a three-dimensional scene simulation system based on 3D Gaussian sputtering of the present invention.

[0047] Figure 3 It is a schematic diagram of a multi-viewing angle image of a real object of the present invention.

[0048] Figure 4 It is a schematic diagram of marking the simulation target according to the present invention.

[0049] Figure 5 Schematic diagram of the triangular mesh model of the present invention.

[0050] Figure 6 It is a schematic diagram of the tetrahedral mesh model of the present invention.

[0051] Figure 7 Schematic diagram of the 3D Gaussian sputtering model of the present invention.

[0052] Figure 8 It is a schematic diagram of the simulation result of deforming the simulation target according to the present invention. DETAILED DESCRIPTION

[0053] The technical solution in the embodiments of the present application has the following overall idea: by marking the simulation targets, the focus is placed on the reconstruction of the simulation targets, while ignoring the interference of background redundant information; by configuring the physical properties, combining implicit integration and finite element methods, the tetrahedral mesh model is controlled to simulate the deformation and motion effects of the simulation targets, fully considering the physical laws behind the changes in the simulation targets; by performing collision detection and response through penalty force or IPC, the collision detection and response problems between multiple reconstructed targets can be well handled, thereby improving the simulation accuracy of the three-dimensional scene.

[0054] Please refer to Figures 1 to 8 As shown, a preferred embodiment of a three-dimensional scene simulation method based on 3D Gaussian sputtering of the present invention comprises the following steps:

[0055] Step S1, obtaining real images of the simulation target from different perspectives, annotating the simulation target in each of the real images to obtain a mask label, and reconstructing an initial point cloud model based on each of the real images, that is, learning the details of the real scene;

[0056] Step S2: reconstructing a triangular mesh model of the simulation target based on the initial point cloud model and the mask label, and expanding each triangular mesh model to obtain a tetrahedral mesh model;

[0057] Step S3, creating a corresponding 3D Gaussian sputtering model based on the triangular mesh model through a 3D Gaussian kernel function;

[0058] The geometric models required by the physical simulation module include tetrahedral mesh model and 3D Gaussian sputtering model. The tetrahedral mesh model can be extracted from the triangular mesh model, while the 3D Gaussian sputtering model needs to match the triangular surface proxy of the triangular mesh model, so the triangular mesh model is reconstructed first.

[0059] Step S4, configuring the physical properties of the simulation target, controlling the tetrahedral mesh model to simulate the deformation and motion effect of the simulation target based on the physical properties through implicit integration and finite element methods, and collecting the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model during the deformation process;

[0060] Step S5, synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient;

[0061] Step S6: Map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

[0062] The step S1 is specifically as follows:

[0063] Acquire physical images of the simulated target at different viewing angles through a camera, for example, acquire 100-300 physical images in RGB format, capture scene details at all angles as much as possible during shooting, and store each of the physical images in real time to a preset image storage path; annotate the simulated target in each of the physical images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path;

[0064] Image features are extracted from each of the physical images through COLMAP, and each of the image features is input into Structure-from-Motion (SfM), thereby reconstructing an initial point cloud model in ply format, and the initial point cloud model is stored in real time to a preset model storage path.

[0065] The step S4 is specifically as follows:

[0066] The physical properties of the simulation target are configured based on the JSON file, wherein the physical properties include at least material type, material parameters, and force conditions; the material parameters include at least Young's modulus, which is used to simulate the morphology of different substances; the force conditions include at least gravitational acceleration, which is used to drive the movement and deformation of the simulation target;

[0067] Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

[0068] That is, the physical grid and preset physical parameters used to solve the simulation equations are output, and the position sequence information (deformation gradient and vertex coordinates) of the physical grid at each time step is output.

[0069] The step S5 is specifically as follows:

[0070] The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

[0071] That is, the deformation gradient is transferred based on the correspondence between the mesh face elements of the tetrahedral mesh model and the triangular mesh model.

[0072] The step S6 is specifically as follows:

[0073] The deformed 3D Gaussian sputtering model is mapped to the tetrahedron mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through the triangular mesh model in combination with penalty force or IPC to interact with different simulation results. That is, the 3D Gaussian sputtering model is bound to the surface of the triangular mesh model to coordinate with the vertex changes of the tetrahedron mesh model and to achieve rendering that restores the texture of the real scene.

[0074] A preferred embodiment of a three-dimensional scene simulation system based on 3D Gaussian sputtering of the present invention includes the following modules:

[0075] The initial point cloud model creation module is used to obtain physical images of the simulation target from different perspectives, annotate the simulation target in each physical image to obtain a mask label, and reconstruct the initial point cloud model based on each physical image, that is, learn the details of the real scene;

[0076] A tetrahedral mesh model creation module, used to reconstruct a triangular mesh model of a simulation target based on the initial point cloud model and the mask label, and expand each triangular mesh model to obtain a tetrahedral mesh model;

[0077] A 3D Gaussian sputtering model creation module, used to create a corresponding 3D Gaussian sputtering model based on the triangular mesh model by using a 3D Gaussian kernel function;

[0078] The geometric models required by the physical simulation module include tetrahedral mesh model and 3D Gaussian sputtering model. The tetrahedral mesh model can be extracted from the triangular mesh model, while the 3D Gaussian sputtering model needs to match the triangular surface proxy of the triangular mesh model, so the triangular mesh model is reconstructed first.

[0079] A tetrahedral mesh model deformation module is used to configure the physical properties of the simulation target, and control the deformation and motion effects of the simulation target based on the physical properties by implicit integration and finite element methods. During the deformation process, deformation gradients and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are collected;

[0080] A 3D Gaussian sputtering model deformation module, used for synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient;

[0081] The simulation rendering module is used to map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

[0082] The initial point cloud model creation module is specifically used for:

[0083] Acquire physical images of the simulated target at different viewing angles through a camera, for example, acquire 100-300 physical images in RGB format, capture scene details at all angles as much as possible during shooting, and store each of the physical images in real time to a preset image storage path; annotate the simulated target in each of the physical images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path;

[0084] Image features are extracted from each of the physical images through COLMAP, and each of the image features is input into Structure-from-Motion (SfM), thereby reconstructing an initial point cloud model in ply format, and the initial point cloud model is stored in real time to a preset model storage path.

[0085] The tetrahedral mesh model deformation module is specifically used for:

[0086] The physical properties of the simulation target are configured based on the JSON file, wherein the physical properties include at least material type, material parameters, and force conditions; the material parameters include at least Young's modulus, which is used to simulate the morphology of different substances; the force conditions include at least gravitational acceleration, which is used to drive the movement and deformation of the simulation target;

[0087] Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

[0088] That is, the physical grid and preset physical parameters used to solve the simulation equations are output, and the position sequence information (deformation gradient and vertex coordinates) of the physical grid at each time step is output.

[0089] The 3D Gaussian sputtering model deformation module is specifically used for:

[0090] The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

[0091] That is, the deformation gradient is transferred based on the correspondence between the mesh face elements of the tetrahedral mesh model and the triangular mesh model.

[0092] The simulation rendering module is specifically used for:

[0093] The deformed 3D Gaussian sputtering model is mapped to the tetrahedron mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through the triangular mesh model in combination with penalty force or IPC to interact with different simulation results. That is, the 3D Gaussian sputtering model is bound to the surface of the triangular mesh model to coordinate with the vertex changes of the tetrahedron mesh model and to achieve rendering that restores the texture of the real scene.

[0094] In summary, the advantages of the present invention are:

[0095] 1. By acquiring real images of the simulation target from different perspectives, the simulation target in each real image is marked to obtain a mask label, and the initial point cloud model is reconstructed based on each real image; then the triangular mesh model of the simulation target is reconstructed based on the initial point cloud model and the mask label, and each triangular mesh model is expanded to obtain a tetrahedral mesh model; then the corresponding 3D Gaussian sputtering model is created based on the triangular mesh model through the 3D Gaussian kernel function; then the physical properties of the simulation target are configured, and the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effect of the simulation target through implicit integration and finite element methods. During the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are collected; then the 3D Gaussian sputtering model is synchronously deformed based on the deformation gradient, and finally The deformed 3D Gaussian sputtering model is mapped to the tetrahedral mesh model based on the vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results; that is, by marking the simulation target, the focus is on the reconstruction of the simulation target and the interference of background redundant information is ignored; the tetrahedral mesh model is controlled through the configured physical properties, combined with implicit integration and finite element methods to simulate the deformation and motion effects of the simulation target, and the physical laws behind the changes of the simulation target are fully considered; collision detection and response are performed through penalty force or IPC, which can well handle the collision detection and response problems between multiple reconstructed targets (simulation results), and ultimately greatly improve the simulation accuracy of the three-dimensional scene.

[0096] 2. Through each time step in the deformation process, the energy equation is solved to obtain the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model, and the deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model to synchronously deform the 3D Gaussian sputtering model, thereby greatly improving the timeliness of the simulation.

[0097] 3. The tetrahedral mesh model is obtained by expanding the triangular mesh model. The tetrahedral mesh model is used as the main model of the simulation. It can be applied to implicit integration and finite element simulation methods to solve the physical motion and deformation equations as accurately as possible while supporting a large time step and ensuring system stability, thereby obtaining realistic simulation results.

[0098] Although the specific implementation modes of the present invention are described above, those skilled in the art should understand that the specific implementation modes described are only illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A three-dimensional scene simulation method based on 3D Gaussian sputtering, characterized in that: The steps include: Step S1, obtaining physical images of the simulation target at different viewing angles, marking the simulation target in each of the physical images to obtain a mask label, and reconstructing an initial point cloud model based on each of the physical images; Step S2: reconstructing a triangular mesh model of the simulation target based on the initial point cloud model and the mask label, and expanding each triangular mesh model to obtain a tetrahedral mesh model; Step S3, creating a corresponding 3D Gaussian sputtering model based on the triangular mesh model through a 3D Gaussian kernel function; Step S4, configuring the physical properties of the simulation target, controlling the tetrahedral mesh model to simulate the deformation and motion effect of the simulation target based on the physical properties through implicit integration and finite element methods, and collecting the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model during the deformation process; Step S5, synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient; Step S6: Map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

2. A three-dimensional scene simulation method based on 3D Gaussian sputtering as claimed in claim 1, characterized in that: The step S1 is specifically as follows: Acquire real images of the simulated target at different viewing angles through a camera, and store each of the real images in real time to a preset image storage path; label the simulated target in each of the real images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path; Image features are extracted from each of the physical object images through COLMAP, and each of the image features is input into Structure-from-Motion to reconstruct an initial point cloud model, and the initial point cloud model is stored in real time to a preset model storage path.

3. The three-dimensional scene simulation method based on 3D Gaussian sputtering according to claim 1, characterized in that: The step S4 is specifically as follows: Configure the physical properties of the simulation target based on the JSON file, where the physical properties include at least material type, material parameters, and stress conditions; Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

4. The three-dimensional scene simulation method based on 3D Gaussian sputtering according to claim 1, characterized in that: The step S5 is specifically as follows: The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

5. The three-dimensional scene simulation method based on 3D Gaussian sputtering according to claim 1, characterized in that: The step S6 is specifically as follows: The deformed 3D Gaussian sputtering model is mapped to a tetrahedral mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results.

6. A three-dimensional scene simulation system based on 3D Gaussian sputtering, characterized in that: Includes the following modules: An initial point cloud model creation module is used to obtain physical images of the simulation target from different perspectives, annotate the simulation target in each physical image to obtain a mask label, and reconstruct the initial point cloud model based on each physical image; A tetrahedral mesh model creation module, used to reconstruct a triangular mesh model of a simulation target based on the initial point cloud model and the mask label, and expand each triangular mesh model to obtain a tetrahedral mesh model; A 3D Gaussian sputtering model creation module, used to create a corresponding 3D Gaussian sputtering model based on the triangular mesh model by using a 3D Gaussian kernel function; A tetrahedral mesh model deformation module is used to configure the physical properties of the simulation target, and control the deformation and motion effects of the simulation target based on the physical properties by implicit integration and finite element methods. During the deformation process, deformation gradients and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are collected; A 3D Gaussian sputtering model deformation module, used for synchronously deforming the 3D Gaussian sputtering model based on the deformation gradient; The simulation rendering module is used to map the deformed 3D Gaussian sputtering model onto a tetrahedral mesh model based on vertex coordinates, and render to obtain a simulation result.

7. A three-dimensional scene simulation system based on 3D Gaussian sputtering as claimed in claim 6, characterized in that: The initial point cloud model creation module is specifically used for: Acquire real images of the simulated target at different viewing angles through a camera, and store each of the real images in real time to a preset image storage path; label the simulated target in each of the real images to obtain a mask label, and store each of the mask labels in real time to a preset label storage path; Image features are extracted from each of the physical object images through COLMAP, and each of the image features is input into Structure-from-Motion to reconstruct an initial point cloud model, and the initial point cloud model is stored in real time to a preset model storage path.

8. A three-dimensional scene simulation system based on 3D Gaussian sputtering as claimed in claim 6, characterized in that: The tetrahedral mesh model deformation module is specifically used for: Configure the physical properties of the simulation target based on the JSON file, where the physical properties include at least material type, material parameters, and stress conditions; Through implicit integration and finite element methods, the tetrahedral mesh model is controlled based on the physical properties to simulate the deformation and motion effects of the simulation target. At each time step in the deformation process, the deformation gradient and vertex coordinates of each tetrahedral mesh in the tetrahedral mesh model are obtained by solving the energy equation.

9. A three-dimensional scene simulation system based on 3D Gaussian sputtering as claimed in claim 6, characterized in that: The 3D Gaussian sputtering model deformation module is specifically used for: The deformation gradient is synchronously transmitted to the 3D Gaussian kernel function corresponding to the triangular mesh model, and the covariance matrix and the center point of the 3D Gaussian kernel function are modified to synchronously deform the 3D Gaussian sputtering model based on the deformation gradient.

10. The three-dimensional scene simulation system based on 3D Gaussian sputtering according to claim 6, characterized in that: The simulation rendering module is specifically used for: The deformed 3D Gaussian sputtering model is mapped to a tetrahedral mesh model based on vertex coordinates, and the simulation result is rendered; collision detection and response are performed between different simulation results through a triangular mesh model combined with penalty force or IPC to interact with different simulation results.

Images