Traffic scene construction method and system based on multi-source data fusion, and storage medium

Abstract

The invention discloses a traffic scene construction method and system based on multi-source data fusion, and a storage medium. The method comprises the steps of obtaining DGM map data, OSM data and vehicle body sensing data of a target area; road space feature information, building space feature information and static and dynamic target space feature information are extracted from the DGM map data, the OSM data and the vehicle body sensing data respectively; respectively constructing a road virtual model, a building virtual model and a static and dynamic target object ply model according to the road space feature information, the building space feature information and the static and dynamic target object space feature information; and carrying out space-time fusion on the road virtual model, the building virtual model and the static and dynamic target object ply model to generate a traffic virtual scene. According to the method, DGM map data, OSM data and vehicle body sensing data are fused, advantage complementation among different data sources is achieved, and high-precision simulation of complex traffic scenes is improved.

Description

Technical Field

[0001] The present invention relates to the field of virtual simulation technology, and in particular to a traffic scene construction method, system and storage medium based on multi-source data fusion. Background Art

[0002] Virtual traffic scene construction technology plays a crucial role in autonomous driving simulation. However, existing virtual traffic scene construction technologies rely on a single data source to construct complex traffic scenarios, severely lacking surrounding environmental or road information. This results in significant deficiencies in the integrity of the constructed scenarios in simulating real-world traffic environments, making them unable to independently meet the requirements for high-precision simulation of complex scenarios. Therefore, how to construct high-precision complex traffic scenes is a key issue that needs to be addressed urgently.

[0003] The above content is only used to assist in understanding the technical solution of the present invention and does not constitute an admission that the above content is prior art. Summary of the Invention The main purpose of the present invention is to provide a traffic scene construction method, system and storage medium based on multi-source data fusion, aiming to solve the technical problem of how to construct complex traffic scenes with high precision.

[0004] To achieve the above objectives, the present invention provides a traffic scene construction method based on multi-source data fusion, the traffic scene construction method based on multi-source data fusion comprising: Obtain DGM map data, OSM data and vehicle body perception data of the target area; Extracting road space feature information, building space feature information, and static and dynamic target space feature information from the DGM map data, the OSM data, and the vehicle body perception data, respectively; Constructing a road virtual model, a building virtual model and a static and dynamic target object ply model respectively according to the road space feature information, the building space feature information and the static and dynamic target object space feature information; The road virtual model, the building virtual model and the static and dynamic target object ply model are temporally and spatially fused to generate a traffic virtual scene.

[0005] Optionally, constructing a road virtual model according to the road space feature information includes: Extracting GPS coordinate points of lane outer contour, GPS coordinate points of lane lines, lane line type, lane line color, road surface arrow mark type and GPS coordinate points of road surface arrows from the road spatial feature information; Generate a lane outline array, a lane line array, and a road arrow array according to the GPS coordinate points of the lane outline, the GPS coordinate points of the lane lines, and the GPS coordinate points of the road arrows; Creating a road network model through a virtual engine based on the lane outer contour array, the lane line array, and the road arrow array; The road network model is rendered according to the lane line type, the lane line color and the road surface arrow mark type to obtain a road virtual model.

[0006] Optionally, generating a lane outer contour array, a lane line array, and a road arrow array according to the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane lines, and the GPS coordinate points of the road arrows, respectively, includes: Perform coordinate conversion on the GPS coordinate points of the lane outer contour, the lane line GPS coordinate points, and the road arrow GPS coordinate points respectively to obtain the lane outer contour coordinate data, lane line coordinate data, and road arrow coordinate data in the UE coordinate system; A lane outer contour array, a lane line array and a road surface arrow array are generated respectively according to the lane outer contour coordinate data, the lane line coordinate data and the road surface arrow coordinate data.

[0007] Optionally, constructing a virtual building model according to the building space feature information includes: Extracting building outer contour coordinate information and building location information from the building space feature information; A building virtual model is constructed through a virtual engine according to the building outer contour coordinate information and the building position information.

[0008] Optionally, constructing a ply model of a static and dynamic target object according to the spatial feature information of the static and dynamic target object includes: Extracting graphic information of static targets and graphic information of dynamic targets from the spatial feature information of the static and dynamic targets, and extracting radar point cloud data of the targets from the vehicle body perception data; Generating a multi-viewing view of a static object and a multi-viewing view of a dynamic object according to the graphic information of the static object and the graphic information of the dynamic object respectively; A ply model of a static and dynamic target is constructed according to the multi-perspective view of the static target, the multi-perspective view of the dynamic target and the radar point cloud data of the target.

[0009] Optionally, constructing a ply model of a static and dynamic target object according to the multi-perspective view of the static target object, the multi-perspective view of the dynamic target object, and the radar point cloud data of the target object includes: Extracting the radar point cloud data of the static target and the radar point cloud data of the dynamic target from the target radar point cloud data; Constructing a ply model of a static target object using 3DGS according to the multi-view views of the static target object and the corresponding radar point cloud data, and constructing a ply model of a dynamic target object using 3DGS according to the multi-view views of the dynamic target object and the corresponding radar point cloud data; A static and dynamic target ply model is generated based on the static target ply model and the dynamic target ply model.

[0010] Optionally, constructing a ply model of the dynamic target object through 3DGS according to the multi-view views of the dynamic target object and the corresponding radar point cloud data includes: Constructing an initial target ply model through 3DGS according to the multi-view view of the dynamic target and the corresponding radar point cloud data; generating a target motion trajectory according to the radar point cloud data of the dynamic target; The center point of the initial target object ply model is aligned with the target object motion trajectory to obtain a dynamic target object ply model.

[0011] Optionally, the spatiotemporal fusion of the road virtual model, the building virtual model and the dynamic target object ply model to generate a traffic virtual scene includes: Unify the UE coordinates corresponding to the road virtual model, the UE coordinates corresponding to the building virtual model, and the Cartesian coordinates corresponding to the static and dynamic target object ply model; An initial virtual scene is created through a 3D graphics engine, and the road virtual model, building virtual model and static and dynamic target object ply model with unified coordinates are imported into the initial virtual scene to generate a traffic virtual scene.

[0012] In addition, to achieve the above-mentioned purpose, the present invention further proposes a traffic scene construction system based on multi-source data fusion, the traffic scene construction system based on multi-source data fusion comprising: Acquisition module, used to obtain DGM map data, OSM data and vehicle body perception data of the target area; a processing module for extracting road space feature information, building space feature information, and static and dynamic target space feature information from the DGM map data, the OSM data, and the vehicle body perception data, respectively; A generation module, configured to construct a road virtual model, a building virtual model, and a static and dynamic target object ply model respectively according to the road space feature information, the building space feature information, and the static and dynamic target object space feature information; The fusion module is used to perform spatiotemporal fusion of the road virtual model, the building virtual model and the static and dynamic target object ply model to generate a traffic virtual scene.

[0013] In addition, to achieve the above-mentioned purpose, the present invention also proposes a traffic scene construction device based on multi-source data fusion, which includes: a memory, a processor, and a traffic scene construction program based on multi-source data fusion stored on the memory and runnable on the processor. The traffic scene construction program based on multi-source data fusion is configured to implement the steps of the traffic scene construction method based on multi-source data fusion as described above.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which a traffic scene construction program based on multi-source data fusion is stored. When the traffic scene construction program based on multi-source data fusion is executed by a processor, the steps of the traffic scene construction method based on multi-source data fusion as described above are implemented.

[0015] The present invention first obtains the target area's DGM map data, OSM data, and vehicle perception data. It then extracts road spatial feature information, building spatial feature information, and static and dynamic target spatial feature information from these data. It then constructs virtual road models, virtual building models, and ply models of static and dynamic targets based on these information. Finally, these virtual road models, virtual building models, and ply models of static and dynamic targets are spatially and temporally fused to generate a virtual traffic scene. By integrating high-precision DGM maps and OSM data, the present invention achieves accurate road information generation and reasonable construction of surrounding buildings. It also utilizes vehicle perception data to supplement traffic elements such as vehicles and pedestrians in real time. The virtual road models, virtual building models, and ply models of static and dynamic targets are spatially and temporally fused to create a highly accurate and complex traffic scene, providing a more realistic simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a structural diagram of a traffic scene construction device based on multi-source data fusion in the hardware operating environment involved in the embodiment of the present invention; Figure 2 This is a flow chart of a first embodiment of a method for constructing a traffic scene based on multi-source data fusion according to the present invention; Figure 3 This is a structural block diagram of the first embodiment of the traffic scene construction system based on multi-source data fusion of the present invention.

[0017] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0018] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] Reference Figure 1 , Figure 1 The diagram is a schematic diagram of a device structure for constructing a traffic scene based on multi-source data fusion in the hardware operating environment involved in the embodiment of the present invention.

[0020] like Figure 1 As shown, the traffic scene construction device based on multi-source data fusion may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display and an input unit such as a keyboard. Optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless Fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk storage device. The memory 1005 may also be a storage system independent of the processor 1001.

[0021] Those skilled in the art will understand that Figure 1 The structure shown in does not constitute a limitation on the traffic scene construction device based on multi-source data fusion, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0022] like Figure 1 As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and a traffic scene construction program based on multi-source data fusion.

[0023] exist Figure 1In the traffic scene construction device based on multi-source data fusion shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the traffic scene construction device based on multi-source data fusion of the present invention can be set in the traffic scene construction device based on multi-source data fusion, and the traffic scene construction device based on multi-source data fusion calls the traffic scene construction program based on multi-source data fusion stored in the memory 1005 through the processor 1001, and executes the traffic scene construction method based on multi-source data fusion provided by the embodiment of the present invention.

[0024] The embodiment of the present invention provides a method for constructing a traffic scene based on multi-source data fusion. Figure 2 , Figure 2 This is a flow chart of the first embodiment of the traffic scene construction method based on multi-source data fusion of the present invention.

[0025] In this embodiment, the traffic scene construction method based on multi-source data fusion includes the following steps: Step S10: Obtain DGM map data, OSM data and vehicle body perception data of the target area.

[0026] It is easy to understand that the execution subject of this embodiment can be a traffic scene construction system based on multi-source data fusion with functions such as data processing, network communication and program running, or it can be other computer devices with similar functions, etc. This embodiment is not limited.

[0027] It should be noted that the target area is the area where the vehicle is currently traveling, and the body perception data is the surrounding environment data collected by the vehicle's onboard sensors.

[0028] Step S20: extracting road space feature information, building space feature information and static and dynamic target space feature information from the DGM map data, the OSM data and the vehicle body perception data respectively.

[0029] In the specific implementation, the Digital Ground Model (DGM) map data and the Open Street Map (OSM) data are parsed respectively, and the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line, the lane line type, the lane line color, the road surface arrow marking type and the GPS coordinate points of the road surface arrow are extracted from the parsed DGM map data, and the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line, the lane line type, the lane line color, the road surface arrow marking type and the GPS coordinate points of the road surface arrow are used as road space feature information; the building outer contour coordinate information and the building location information are extracted from the parsed OSM data, and the building outer contour coordinate information and the building location information are used as building space feature information.

[0030] The graphic information of static targets and the corresponding radar point cloud data, as well as the graphic information of dynamic targets and the corresponding radar point cloud data, are extracted from the vehicle body perception data, and the graphic information of static targets and the corresponding radar point cloud data, as well as the graphic information of dynamic targets and the corresponding radar point cloud data, are used as the spatial feature information of static and dynamic targets.

[0031] Static targets can be static objects on the road, such as trees, trash cans, etc. Dynamic targets include pedestrians and other vehicles.

[0032] Step S30: constructing a road virtual model, a building virtual model, and a static and dynamic target ply model respectively according to the road space feature information, the building space feature information, and the static and dynamic target space feature information.

[0033] Furthermore, a processing method for constructing a road virtual model based on road spatial feature information is as follows: extracting the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line, the lane line type, the lane line color, the road surface arrow marking type and the GPS coordinate points of the road surface arrow from the road spatial feature information; generating a lane outer contour array, a lane line array and a road surface arrow array based on the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line and the GPS coordinate points of the road surface arrow respectively; creating a road network model through a virtual engine based on the lane outer contour array, the lane line array and the road surface arrow array; rendering the road network model based on the lane line type, the lane line color and the road surface arrow marking type to obtain a road virtual model.

[0034] The processing method for generating the lane outer contour array, lane line array and road surface arrow array respectively according to the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line and the GPS coordinate points of the road surface arrow is as follows: coordinate conversion is performed on the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line and the GPS coordinate points of the road surface arrow respectively to obtain the lane outer contour coordinate data, lane line coordinate data and road surface arrow coordinate data in the UE coordinate system; and the lane outer contour array, lane line array and road surface arrow array are generated respectively according to the lane outer contour coordinate data, lane line coordinate data and road surface arrow coordinate data.

[0035] In the specific implementation, the GPS coordinate points of the lane outer contour, the lane line GPS coordinate points, and the road arrow GPS coordinate points are converted into the UE coordinate system using the Sanson-Flamsteed projection, becoming the familiar xyz data, and obtaining the lane outer contour coordinate data, lane line coordinate data, and road arrow coordinate data in the UE coordinate system. The lane outer contour coordinate data is then converted into an array and then into a spline through the UE blueprint. The spline is further converted into a poly path using the blueprint, and polygons (i.e., triangular meshes) are generated based on the poly path using the blueprint. The lane line coordinate data is the trajectory of the lane line. The UE blueprint converts the data into an array and then into a spline. UE PCG is used to place rectangular lane line models along the spline at a certain interval. A road arrow array is generated based on the road arrow coordinate data, and the UE's spawn function is used to place the road arrow model at the specified position (the model is a model made in the virtual engine (3ds max) and imported into UE). Finally, the blueprint is used to assign road materials to the polygons, and the materials and textures are rendered according to the lane line and traffic sign properties to obtain a virtual road model.

[0036] Furthermore, the processing method for constructing a building virtual model based on the building space feature information is to extract the building outer contour coordinate information and the building position information from the building space feature information; and construct the building virtual model through a virtual engine based on the building outer contour coordinate information and the building position information.

[0037] It should be noted that the construction method of the building virtual model is the same as that of the road virtual model.

[0038] Furthermore, a processing method for constructing a ply model of static and dynamic targets based on the spatial feature information of static and dynamic targets is to extract the graphic information of the static target and the graphic information of the dynamic target from the spatial feature information of the static and dynamic targets, and extract the radar point cloud data of the target from the vehicle body perception data; generate a multi-perspective view of the static target and a multi-perspective view of the dynamic target according to the graphic information of the static target and the graphic information of the dynamic target respectively; and construct a ply model of the static and dynamic targets based on the multi-perspective view of the static target, the multi-perspective view of the dynamic target and the radar point cloud data of the target.

[0039] The processing method for constructing a ply model of a static and dynamic target according to the multi-view views of a static target, the multi-view views of a dynamic target and the radar point cloud data of the target is as follows: extracting the radar point cloud data of the static target and the radar point cloud data of the dynamic target from the radar point cloud data of the target; constructing a ply model of the static target according to the multi-view views of the static target and the corresponding radar point cloud data through three-dimensional Gaussian scattering (3D Gaussian Splatting 3DGS), and constructing a ply model of the dynamic target according to the multi-view views of the dynamic target and the corresponding radar point cloud data through 3DGS; and generating a ply model of the static and dynamic target based on the static target ply model and the dynamic target ply model.

[0040] In the specific implementation, the DeepLab image segmentation model is used to identify and crop the pixels belonging to the target object in the graphic information, and then import them into the Large Gaussian Reconstruction Model for Efficient 3D Reconstruction and Generation (GRM) or the Large Multi-View Gaussian Model (LGM) algorithm to generate three views (i.e., multi-view views of static targets and multi-view views of dynamic targets).

[0041] Furthermore, a processing method for constructing a ply model of a dynamic target object through 3DGS according to the multi-perspective views of the dynamic target object and the corresponding radar point cloud data is as follows: constructing an initial ply model of the dynamic target object through 3DGS according to the multi-perspective views of the dynamic target object and the corresponding radar point cloud data; generating a motion trajectory of the target object according to the radar point cloud data of the dynamic target object; and aligning the center point of the initial ply model of the target object with the motion trajectory of the target object to obtain the ply model of the dynamic target object.

[0042] In this embodiment, the radar point cloud data comes from the on-board radar of the autonomous vehicle (i.e., the ego vehicle) and is usually stored in a point cloud file format, including the distance information of the target vehicle from the ego vehicle and timestamp information. The trajectory information of the target vehicle (i.e., the target object's motion trajectory) can be obtained through the GPS coordinates of the ego vehicle according to the timestamp, and the center point of the initial target object ply model is aligned with the target object's motion trajectory to obtain a dynamic target object ply model, thereby realizing dynamic supplementation and real-time updating of the scene.

[0043] Step S40: performing spatiotemporal fusion of the road virtual model, the building virtual model, and the static and dynamic target object ply models to generate a traffic virtual scene.

[0044] Furthermore, the UE coordinates corresponding to the road virtual model, the UE coordinates corresponding to the building virtual model, and the Cartesian coordinates corresponding to the static and dynamic target object ply model are unified; an initial virtual scene is created through a 3D graphics engine, and the road virtual model, building virtual model, and static and dynamic target object ply model after coordinate unification are imported into the initial virtual scene to generate a traffic virtual scene, realize the movement of dynamic objects on the generated road, and complete the construction of a virtual scene combining dynamic and static elements.

[0045] In this embodiment, DGM map data, OSM data, and vehicle perception data for the target area are first acquired. Then, spatial feature information for roads, buildings, and static and dynamic objects is extracted from these data. A virtual road model, a virtual building model, and a ply model of static and dynamic objects are constructed based on these spatial feature information. Finally, these virtual road models, virtual building models, and ply models of static and dynamic objects are spatially fused to generate a virtual traffic scene. This embodiment achieves accurate road information generation and reasonable construction of surrounding buildings by fusing DGM high-precision maps and OSM data. Vehicle perception data is used to supplement traffic elements such as vehicles and pedestrians in real time. The virtual road models, virtual building models, and ply models of static and dynamic objects are spatially fused to create a highly accurate and complex traffic scene, providing a more realistic simulation.

[0046] Reference Figure 3 , Figure 3 This is a structural block diagram of the first embodiment of the traffic scene construction system based on multi-source data fusion of the present invention.

[0047] like Figure 3 As shown, the traffic scene construction system based on multi-source data fusion proposed in an embodiment of the present invention includes: Acquisition module 3001, used to acquire DGM map data, OSM data and vehicle body perception data of the target area; The processing module 3002 is used to extract road spatial feature information, building spatial feature information, and static and dynamic object spatial feature information from the DGM map data, the OSM data, and the vehicle body perception data respectively; A generation module 3003 is configured to construct a road virtual model, a building virtual model, and a static and dynamic target object ply model according to the road space feature information, the building space feature information, and the static and dynamic target object space feature information; The fusion module 3004 is used to perform spatiotemporal fusion of the road virtual model, the building virtual model and the static and dynamic target object ply model to generate a traffic virtual scene.

[0048] Other embodiments or specific implementations of the traffic scene construction system based on multi-source data fusion of the present invention can refer to the above-mentioned method embodiments and will not be repeated here.

[0049] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or system comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or system. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or system comprising the element.

[0050] The serial numbers of the above embodiments of the present invention are for description only and do not represent the advantages or disadvantages of the embodiments.

[0051] Through the description of the above embodiments, those skilled in the art will clearly understand that the above-mentioned embodiments and methods can be implemented by means of software plus the necessary general-purpose hardware platform. Of course, hardware can also be used, but in many cases the former is a more preferred embodiment. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory / random access memory, magnetic disk, or optical disk) and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in various embodiments of the present invention.

[0052] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A traffic scene construction method based on multi-source data fusion, characterized in that: The method comprises the following steps: Obtain DGM map data, OSM data and vehicle body perception data of the target area; Extracting road space feature information, building space feature information, and static and dynamic target space feature information from the DGM map data, the OSM data, and the vehicle body perception data, respectively; Constructing a road virtual model, a building virtual model and a static and dynamic target object ply model respectively according to the road space feature information, the building space feature information and the static and dynamic target object space feature information; The road virtual model, the building virtual model and the static and dynamic target object ply model are temporally and spatially fused to generate a traffic virtual scene.

2. The method according to claim 1, wherein The step of constructing a road virtual model according to the road space feature information includes: Extracting GPS coordinate points of lane outer contour, GPS coordinate points of lane lines, lane line type, lane line color, road surface arrow mark type and GPS coordinate points of road surface arrows from the road spatial feature information; Generate a lane outline array, a lane line array, and a road arrow array according to the GPS coordinate points of the lane outline, the GPS coordinate points of the lane lines, and the GPS coordinate points of the road arrows; Creating a road network model through a virtual engine based on the lane outer contour array, the lane line array, and the road arrow array; The road network model is rendered according to the lane line type, the lane line color and the road surface arrow mark type to obtain a road virtual model.

3. The method according to claim 2, wherein The method of generating a lane outer contour array, a lane line array, and a road arrow array according to the GPS coordinate points of the lane outer contour, the GPS coordinate points of the lane line, and the GPS coordinate points of the road arrow, respectively, includes: Perform coordinate conversion on the GPS coordinate points of the lane outer contour, the lane line GPS coordinate points, and the road arrow GPS coordinate points respectively to obtain the lane outer contour coordinate data, lane line coordinate data, and road arrow coordinate data in the UE coordinate system; A lane outer contour array, a lane line array and a road surface arrow array are generated respectively according to the lane outer contour coordinate data, the lane line coordinate data and the road surface arrow coordinate data.

4. The method according to claim 1, wherein The step of constructing a virtual building model according to the building space characteristic information includes: Extracting building outer contour coordinate information and building location information from the building space feature information; A building virtual model is constructed through a virtual engine according to the building outer contour coordinate information and the building position information.

5. The method according to claim 1, wherein The step of constructing a ply model of a static and dynamic target object according to the spatial feature information of the static and dynamic target object comprises: Extracting graphic information of static targets and graphic information of dynamic targets from the spatial feature information of the static and dynamic targets, and extracting radar point cloud data of the targets from the vehicle body perception data; Generating a multi-viewing view of a static object and a multi-viewing view of a dynamic object according to the graphic information of the static object and the graphic information of the dynamic object respectively; A ply model of a static and dynamic target is constructed according to the multi-perspective view of the static target, the multi-perspective view of the dynamic target and the radar point cloud data of the target.

6. The method according to claim 5, wherein The constructing of a ply model of a static and dynamic target object according to the multi-view view of the static target object, the multi-view view of the dynamic target object and the radar point cloud data of the target object includes: Extracting the radar point cloud data of the static target and the radar point cloud data of the dynamic target from the target radar point cloud data; Constructing a ply model of a static target object using 3DGS according to the multi-view views of the static target object and the corresponding radar point cloud data, and constructing a ply model of a dynamic target object using 3DGS according to the multi-view views of the dynamic target object and the corresponding radar point cloud data; A static and dynamic target ply model is generated based on the static target ply model and the dynamic target ply model.

7. The method according to claim 6, wherein The constructing of a dynamic target ply model by 3DGS according to the multi-view view of the dynamic target and the corresponding radar point cloud data includes: Constructing an initial target ply model through 3DGS according to the multi-view view of the dynamic target and the corresponding radar point cloud data; generating a target motion trajectory according to the radar point cloud data of the dynamic target; The center point of the initial target object ply model is aligned with the target object motion trajectory to obtain a dynamic target object ply model.

8. The method according to any one of claims 1 to 7, wherein: The step of performing spatiotemporal fusion of the road virtual model, the building virtual model, and the dynamic target object ply model to generate a traffic virtual scene includes: Unify the UE coordinates corresponding to the road virtual model, the UE coordinates corresponding to the building virtual model, and the Cartesian coordinates corresponding to the static and dynamic target object ply model; An initial virtual scene is created through a 3D graphics engine, and the road virtual model, building virtual model and static and dynamic target object ply model with unified coordinates are imported into the initial virtual scene to generate a traffic virtual scene.

9. A traffic scene construction system based on multi-source data fusion, characterized in that: The system comprises: Acquisition module, used to obtain DGM map data, OSM data and vehicle body perception data of the target area; a processing module for extracting road space feature information, building space feature information, and static and dynamic target space feature information from the DGM map data, the OSM data, and the vehicle body perception data, respectively; A generation module, configured to construct a road virtual model, a building virtual model, and a static and dynamic target object ply model respectively according to the road space feature information, the building space feature information, and the static and dynamic target object space feature information; The fusion module is used to perform spatiotemporal fusion of the road virtual model, the building virtual model and the static and dynamic target object ply model to generate a traffic virtual scene.

10. A storage medium, characterized in that: The storage medium stores a traffic scene construction program based on multi-source data fusion. When the traffic scene construction program based on multi-source data fusion is executed by the processor, the steps of the traffic scene construction method based on multi-source data fusion as described in any one of claims 1 to 8 are implemented.

Images