Virtual-real combination-based rail transit perception data simulation method and device

Abstract

The embodiment of the invention provides a rail transit perception data simulation method and device based on virtuality and reality combination, and the method comprises the steps: determining rail transit scene components, obtaining real line data, carrying out the modeling based on the rail transit scene components and the real line data, and carrying out the simulation of the rail transit perception data. Establishing a rail transit virtual simulation system with weather parameters; creating a weather manager, and managing and controlling weather parameters of the rail transit virtual simulation system through the weather manager; simulation image information sent by a simulation industrial camera and / or simulation sensor information sent by a simulation sensor are / is obtained, and based on the rail transit virtual simulation system, visual operation is carried out according to the simulation image information and the simulation sensor information. Visual display of different industrial cameras and sensors in a rail transit scene is realized. The embodiment of the invention has a promotion effect on popularization and application of a future travel perception system.

Description

Technical Field

[0001] This document relates to the field of computer technology, and in particular to a method and device for simulating rail transit perception data based on a combination of virtual and real. Background Art

[0002] Perception simulation testing can simulate scenes that are difficult to reproduce in actual projects, and simulate scene sensor data to provide a richer data set for the algorithm, improving the perception effect of the perception system in actual applications. Testing and verification of perception capabilities can provide higher quality safety assurance for the application of autonomous driving perception systems, but there are no mature and large-scale application cases in the field of urban rail transit.

[0003] Domestic and foreign researchers have conducted extensive research on test integration in the automotive autonomous driving industry, but rarely on perception simulation in the rail transit field. The complexity, diversity, and high dynamics of rail transit scenarios pose huge challenges to the development and verification of perception systems, such as inclement weather scenarios such as rain, snow, and fog. Traditional perception testing methods mainly rely on data collection and experiments in actual environments. However, this approach is not only costly and time-consuming, but also difficult to cover all possible scenarios and boundary conditions. Summary of the invention

[0004] The purpose of the present invention is to provide a method and device for simulating rail transit perception data based on a combination of virtual and real, aiming to solve the above-mentioned problems in the prior art.

[0005] The present invention provides a rail transit perception data simulation method based on virtual-real combination, comprising:

[0006] Determine the rail transit scene components, obtain real line data, perform modeling based on the rail transit scene components and the real line data, and establish a rail transit virtual simulation system with weather parameters;

[0007] Creating a weather manager, and managing and controlling weather parameters of the rail transit virtual simulation system through the weather manager;

[0008] Acquire simulated image information sent by simulated industrial cameras and / or simulated sensor information sent by simulated sensors, and based on the rail transit virtual simulation system, perform visualization operations according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in rail transit scenes.

[0009] The present invention provides a rail transit perception data simulation device based on virtual-real combination, comprising:

[0010] A modeling module, used to determine the components of a rail transit scene, obtain real line data, perform modeling based on the components of the rail transit scene and the real line data, and establish a rail transit virtual simulation system with weather parameters;

[0011] A weather module, used to create a weather manager, through which weather parameters of the rail transit virtual simulation system are managed and controlled;

[0012] The visualization module is used to obtain simulated image information sent by simulated industrial cameras and / or simulated sensor information sent by simulated sensors. Based on the rail transit virtual simulation system, visualization operations are performed according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in rail transit scenes.

[0013] An embodiment of the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the steps of the above-mentioned rail transit perception data simulation method based on the combination of virtual and real are implemented.

[0014] An embodiment of the present invention further provides a computer-readable storage medium, on which a program for implementing information transmission is stored. When the program is executed by a processor, the steps of the above-mentioned rail transit perception data simulation method based on the combination of virtual and real are implemented.

[0015] By adopting the embodiment of the present invention, starting from the perception simulation test method in the field of rail transit, the simulation scene construction and perception data simulation method based on the open source UE engine and Carla tool are innovatively studied. The test line scene is modeled through the virtual-reality fusion technology, and the sensor simulation data is simulated according to the Carla sensor simulation model. Finally, it is preliminarily tested and verified in the BEV perception algorithm of the adjacent track train operation scenario. The technical solution of the embodiment of the present invention will promote the promotion and application of future line perception systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate one or more embodiments of this specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0017] Figure 1 is a flow chart of a rail transit perception data simulation method based on virtual-real combination according to an embodiment of the present invention;

[0018] Figure 2 It is a schematic diagram of the simulation test platform system architecture of an embodiment of the present invention, when constructing a rail transit perception test scenario;

[0019] Figure 3 It is a schematic diagram of the test line scene modeling effect based on UE4 in an embodiment of the present invention;

[0020] Figure 4 is a schematic diagram of the Carla built-in simulation effect of an embodiment of the present invention;

[0021] Figure 5 is a schematic diagram of the output effect of the virtual sensor according to an embodiment of the present invention;

[0022] Figure 6 is a schematic diagram of the effect of a train entering and exiting a tunnel according to an embodiment of the present invention;

[0023] Figure 7 is a schematic diagram of the simulation effect of an embodiment of the present invention;

[0024] Figure 8 is a schematic diagram of a pitch angle deviation of an embodiment of the present invention;

[0025] Fig. 9 is a schematic diagram of a rail transit perception data simulation device based on virtual-real combination according to an embodiment of the present invention;

[0026] Fig.10 is a schematic diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the following will be combined with the drawings in one or more embodiments of this specification to clearly and completely describe the technical solutions in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0028] Method Embodiment

[0029] According to an embodiment of the present invention, a rail transit perception data simulation method based on virtual-real combination is provided. Figure 1 is a flow chart of a rail transit perception data simulation method based on virtual-real combination according to an embodiment of the present invention, such as Figure 1 As shown, the rail transit perception data simulation method based on virtual-real combination according to an embodiment of the present invention specifically includes:

[0030] Step S101, determine the rail transit scene components, obtain real line data, model based on the rail transit scene components and the real line data, and establish a rail transit virtual simulation system with weather parameters; wherein the rail transit scene components specifically include: basic element layer, weather element layer, lighting element layer, section element layer and working condition element layer. Step S101 specifically includes:

[0031] Specifically include:

[0032] UE4 is used to model the rail transit scene components and the real line data. The height, undulation and details of the terrain in the rail transit scene are accurately restored by importing GIS data and using Landscape tools. The three-dimensional models of tracks, switches and sleepers are refined by using 3D modeling software. The Spline tool in UE4 is used for precise layout to ensure that the geometric parameters and operating characteristics of the rail system are consistent with the real environment.

[0033] Combined with UE4's blueprint system, dynamic interactive functions are given to each 3D model;

[0034] Seamlessly integrate tunnel and building structures with terrain through custom modeling, and set dynamic weather and lighting conditions by integrating UE4's weather module and real-time lighting functions;

[0035] Manage computing resources through the use of LOD systems, light maps, and instanced materials;

[0036] Managing virtual scenes and setting data generation rules so that the rail transit virtual simulation system can provide different test scenes and output data;

[0037] By using the UE4 Unreal Engine and visually modeling the three-dimensional model in a virtual environment, a rail transit virtual simulation system with weather parameters was constructed.

[0038] Step S102, creating a weather manager, and managing and controlling the weather parameters of the rail transit virtual simulation system through the weather manager; specifically including:

[0039] A weather manager is created through Carla-UE4, corresponding weather parameters are input through the weather manager, weather conditions and model inputs are adjusted to manage and randomly generate weather conditions, and weather parameters of the rail transit virtual simulation system are managed and controlled.

[0040] Step S103, obtaining simulated image information sent by a simulated industrial camera and / or simulated sensor information sent by a simulated sensor, and performing visualization operations based on the simulated image information and the simulated sensor information based on the rail transit virtual simulation system to achieve visualization of different industrial cameras and sensors in rail transit scenes. Specifically including:

[0041] Acquire an industrial camera and / or sensor attached to a fixed position of the vehicle body with 6-DOF simulated by the Carla sensor simulation model, acquire simulated image information, and / or simulated sensor information, and perform visualization operations based on the simulated image information and the simulated sensor information based on the rail transit virtual simulation system to display RGB, depth, semantic and event information, and key railway feature information, wherein the key railway feature information includes: trains, signals, sleepers and gantry, wherein the generation of sensor high-simulation numbers specifically includes: simulating sensor data through a sensor simulation simulator to generate sensor raw data, and based on a sensor correction model, adding noise, data attenuation and / or data distortion to the sensor raw data according to actual needs to obtain corrected sensor high-simulation data. Specifically, since the sensor data simulated by the sensor simulator is only basic simple data, it cannot represent various complex weather conditions, working conditions and other scenes. Therefore, it is necessary to post-process the raw sensor data. Based on the sensor correction model, according to the actual weather or working conditions and other scenes required, the raw sensor data is processed to add noise, data attenuation and / or data distortion to obtain the corrected sensor high-simulation data. The following is an example of adding noise: the noise distribution model provided in the Carla platform is a noise function based on Gaussian distribution, and the noise model based on mathematical statistics cannot fit the noise situation well under the objective conditions of high-speed motion and multiple sources of interference. The "OpenCVLens Distortion" plug-in is used to simulate and correct lens distortion, and the noise distribution is simulated based on the GAN network.

[0042] The above technical solution of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings.

[0043] First, the construction of the rail transit perception virtual test environment based on UE4 is explained in detail: Modeling the rail transit scene requires analyzing the elements of the scene system. Rail transit is different from autonomous vehicles, and its operating environment is relatively fixed. In the process of scene construction, the first task is to clarify and determine the various components of the scene. By reasonably classifying these elements, scene-based autonomous driving tests can be implemented more effectively. Referring to the construction logic of the automobile element set, special consideration is given to the unique elements of rail transit, and finally the scene elements are systematically divided into five levels. In order to show these classifications more specifically, the classification of some scene elements is listed in Table 1.

[0044] Table 1

[0045]

[0046] In real environments, scene elements are extremely rich and diverse, and it is difficult to exhaust all possibilities. Therefore, Table 1 is an inductive summary of typical scene elements rather than a comprehensive enumeration.

[0047] The combined construction method of railway scenes, as shown in Table 1, aims to create a diverse virtual railway environment by integrating hierarchical classic railway elements with virtual assets. The railway virtual scene is composed of multi-level elements. The left side of Table 1 shows the multi-level structure of scene elements in detail. These levels are superimposed on each other to jointly construct a complex railway scene.

[0048] Figure 2 is a schematic diagram of the simulation test platform system architecture of an embodiment of the present invention. When constructing a rail transit perception test scenario, Figure 2 As shown in the figure, UE4, with its powerful rendering and physics engine capabilities, has become the core tool for building a simulation environment. First of all, by obtaining data from real lines, including terrain, track layout, and the location of related facilities, it is possible to ensure a high degree of consistency between the virtual environment and the actual scene. The import of GIS data and the application of Landscape tools enable the accurate restoration of the height, undulations, and details of the terrain, laying the foundation for the virtual scene. Models of tracks, switches, sleepers, etc. are finely modeled using 3D modeling software, and accurately arranged using the Spline tool in UE4 to ensure that the geometric parameters and operating characteristics of the track system are consistent with the real environment.

[0049] At the same time, the models of key facilities such as trains, signals, and tunnels must not only have high fidelity, but also be combined with UE4's blueprint system to give them dynamic interactive functions to simulate behaviors in real operations. Structures such as tunnels and buildings are seamlessly integrated with the terrain through custom modeling to ensure the integrity and authenticity of the virtual environment. The dynamic weather system and lighting conditions are set up by integrating UE4's weather module and real-time lighting functions to simulate a variety of environmental conditions, further enhancing the complexity and authenticity of the test scenarios.

[0050] On this basis, scene optimization and performance debugging are crucial. Through the application of LOD (Level of Details) system, lightmaps and instanced materials, computing resources can be effectively managed to ensure the efficient operation of virtual scenes on different hardware platforms. Finally, through the management of scenes and the setting of data generation rules, UE4 can provide rich test scenes and data outputs, providing a comprehensive and efficient simulation environment for the development and verification of rail transit perception systems. This environment can not only simulate complex rail transit conditions, but also ensure the robustness and reliability of the perception system in practical applications through repeated debugging and optimization.

[0051] By using the UE4 Unreal Engine and visually modeling the 3D model in a virtual environment, a rail transit virtual simulation system is constructed. This framework not only has a high degree of authenticity and immersion, but also has good scalability and customizability. Figure 3 The test line scene modeling effect based on UE4 is demonstrated.

[0052] The perception data simulation based on Carla specifically includes the following processing: In Carla-UE4, a weather manager is created to be responsible for the management and random generation of weather conditions. Figure 4 Schematic diagram of the Carla built-in simulation effect of an embodiment of the present invention. Figure 4 As shown, the weather manager is used in the loop to call the set_random_weather() method in the program loop to randomly change the weather. By inputting the corresponding weather parameters (including quantitative values ​​of engine time, sun altitude angle, weather conditions, etc.), and further tuning the weather conditions and model inputs, the generated simulated weather can be ensured to meet the environmental requirements.

[0053] Currently, industrial cameras (with adjustable resolution, field of view, etc.) can be deployed, including RGB cameras, depth cameras, semantic cameras, and event cameras, as well as typical sensors such as LiDAR (with adjustable scanning frequency and detection distance), IMU, and GNSS. It is possible to attach sensors to fixed positions on the vehicle body with 6-DOF and enable sensors to record scene data in parallel. By adjusting parameters to deploy sensors, the visualization of different sensor information in railway scenes is achieved. RGB, depth, semantic, and event information are displayed, as well as key railway features, such as trains, signals, sleepers, gantries, etc. Figure 5 As shown, Figure 5 It is a schematic diagram of the output effect of the virtual sensor according to an embodiment of the present invention.

[0054] Figure 6 Schematic diagram of the effect of a train entering and exiting a tunnel according to an embodiment of the present invention. Figure 6 The figure shows in detail the whole process of a train running at a speed of 120km / h passing through a tunnel in 0-25 seconds in a virtual simulation scenario. This scenario can better simulate the dynamic behavior and light and dark changes of a train passing through a tunnel at a specific speed and time period.

[0055] Figure 7 Schematic diagram of the simulation effect of an embodiment of the present invention. Figure 7 As shown in the figure, it is proposed to simulate the running scene of adjacent track trains. In this scene, when the distance between trains is close, there are fewer radar return points, so this scene is simulated and verified. BEVDet (Bird's EyeView Detection) is a multi-camera perception algorithm based on bird's-eye view, mainly used for 3D target detection in autonomous driving scenarios. It converts the 2D image data of multiple cameras into a unified BEV space for processing, thereby improving scene understanding ability and target detection accuracy.

[0056] Figure 7 The simulation modeling effect of the adjacent track train scene in the tunnel environment is demonstrated. The train length is 100m, the camera is tilted downward at an angle of 8 degrees, and is about 3.5m from the ground. The image resolution is set to 1280*720. Starting from the origin of the train head, the adjacent track train is simulated to move to a range of 500m to collect simulation images and point cloud data at a fixed frame rate. At the same time, a rectangular annotation box is set for the adjacent track train in each frame image, and the 8 vertex coordinates of the train are annotated. Each frame image corresponds to a annotation file. A total of 1,000 images and 1,000 frames of point clouds are simulated, and point clouds are used for deep supervision training.

[0057] Figure 8 Schematic diagram of pitch angle deviation of an embodiment of the present invention. Figure 8As shown in the figure, the labeling problem caused by adding the pitch angle of the sensor is shown. Due to the downward tilt of 8 degrees and the length of the vehicle of 100m, the coordinates in the z-axis direction will have a deviation of about 100*sin 8≈14m. After adding simulated point cloud data for deep supervision, the current experimental results can basically be successfully identified within 200 meters, which reflects the initial prospects of simulated data in the field of rail transit.

[0058] In summary, the technical solution of the embodiment of the present invention adopts a perception data enhancement and testing technology based on the combination of virtual and real. This technology combines the virtual environment with real data to provide an efficient and low-cost testing environment while ensuring the adaptability and reliability of the perception system in the real environment. Especially in the field of rail transit, through the comprehensive application of virtual test environment and perception data simulation, it is possible to test scenes that are difficult to reproduce, significantly improving the development efficiency of the perception system.

[0059] The embodiment of the present invention draws on the technology of the autonomous driving industry to construct a rail transit scenario system, builds a simulation environment based on UE4 (UnrealEngine 4), simulates sensor perception data based on Carla, simulates complex working condition data, and uses a current BEV (Bird's Eye View) algorithm test case to summarize the experimental situation in stages, paving the way for more abundant simulation applications in the future.

[0060] The embodiment of the present invention combines the actual railway field perception production needs, models the rail transit scene based on the virtual-real combination method, simulates the multi-sensor data through tools such as Carla, expands the diversity of simulation data, and finally tests and verifies a certain scene to prove its effect.

[0061] Device Example 1

[0062] According to an embodiment of the present invention, a rail transit perception data simulation device based on virtual-real combination is provided. Fig. 9 is a schematic diagram of a rail transit perception data simulation device based on virtual-real combination according to an embodiment of the present invention, such as Fig. 9 As shown, the rail transit perception data simulation device based on virtual-real combination according to an embodiment of the present invention specifically includes:

[0063] The modeling module 90 is used to determine the components of the rail transit scene, obtain real line data, perform modeling based on the components of the rail transit scene and the real line data, and establish a rail transit virtual simulation system with weather parameters; the components of the rail transit scene specifically include: basic element layer, weather element layer, lighting element layer, section element layer and working condition element layer. The modeling module 90 is specifically used to:

[0064] UE4 is used to model the rail transit scene components and the real line data. The height, undulation and details of the terrain in the rail transit scene are accurately restored by importing GIS data and using Landscape tools. The three-dimensional models of tracks, switches and sleepers are refined by using 3D modeling software. The Spline tool in UE4 is used for precise layout to ensure that the geometric parameters and operating characteristics of the rail system are consistent with the real environment.

[0065] Combined with UE4's blueprint system, dynamic interactive functions are given to each 3D model;

[0066] Seamlessly integrate tunnel and building structures with terrain through custom modeling, and set dynamic weather and lighting conditions by integrating UE4's weather module and real-time lighting functions;

[0067] Manage computing resources through the use of LOD systems, light maps, and instanced materials;

[0068] Managing virtual scenes and setting data generation rules so that the rail transit virtual simulation system can provide different test scenes and output data;

[0069] Using the UE4 Unreal Engine, a rail transit virtual simulation system was constructed by visually modeling and designing the 3D model in a virtual environment;

[0070] The weather module 92 is used to create a weather manager, and manage and control the weather parameters of the rail transit virtual simulation system through the weather manager; the weather module 92 is specifically used to:

[0071] Create a weather manager through Carla-UE4, input corresponding weather parameters through the weather manager, adjust weather conditions and model inputs to manage and randomly generate weather conditions, and manage and control weather parameters of the rail transit virtual simulation system;

[0072] The visualization module 94 is used to obtain the simulated image information sent by the simulated industrial camera and / or the simulated sensor information sent by the simulated sensor, and based on the rail transit virtual simulation system, perform visualization operations according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in the rail transit scene. The visualization module 94 is specifically used to:

[0073] Acquire an industrial camera and / or sensor attached to a fixed position of the vehicle body with 6-DOF simulated by the Carla sensor simulation model, acquire simulated image information, and / or simulated sensor information, and perform visualization operations based on the simulated image information and the simulated sensor information based on the rail transit virtual simulation system to display RGB, depth, semantic and event information, and key railway feature information, wherein the key railway feature information includes: trains, signals, sleepers and gantry, wherein the generation of sensor high-simulation numbers specifically includes: simulating sensor data through a sensor simulation simulator to generate sensor raw data, and based on a sensor correction model, adding noise, data attenuation and / or data distortion to the sensor raw data according to actual needs to obtain corrected sensor high-simulation data.

[0074] The embodiment of the present invention is a device embodiment corresponding to the above method embodiment. The specific operations of each module can be understood by referring to the description of the method embodiment, which will not be repeated here.

[0075] Device Example 2

[0076] An embodiment of the present invention provides an electronic device, such as Fig.10 As shown, it includes: a memory 100, a processor 102, and a computer program stored in the memory 100 and executable on the processor 102, and when the computer program is executed by the processor 102, the steps described in the method embodiment are implemented.

[0077] Device Example 3

[0078] An embodiment of the present invention provides a computer-readable storage medium, on which a program for implementing information transmission is stored. When the program is executed by the processor 102, the steps described in the method embodiment are implemented.

[0079] The computer-readable storage medium in this embodiment includes, but is not limited to, ROM, RAM, magnetic disk or optical disk, etc.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A rail transit perception data simulation method based on virtual-real combination, characterized in that: include: Determine the rail transit scene components, obtain real line data, perform modeling based on the rail transit scene components and the real line data, and establish a rail transit virtual simulation system with weather parameters; Creating a weather manager, and managing and controlling weather parameters of the rail transit virtual simulation system through the weather manager; Acquire simulated image information sent by simulated industrial cameras and / or simulated sensor information sent by simulated sensors, and based on the rail transit virtual simulation system, perform visualization operations according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in rail transit scenes.

2. The method according to claim 1, characterized in that The rail transit scene components specifically include: a basic element layer, a weather element layer, a lighting element layer, a road section element layer and a working condition element layer.

3. The method according to claim 1, characterized in that Modeling based on the rail transit scene components and the real line data to establish a rail transit virtual simulation system with weather parameters specifically includes: UE4 is used to model the rail transit scene components and the real line data. The height, undulation and details of the terrain in the rail transit scene are accurately restored by importing GIS data and using Landscape tools. The three-dimensional models of tracks, switches and sleepers are refined by using 3D modeling software. The Spline tool in UE4 is used for precise layout to ensure that the geometric parameters and operating characteristics of the rail system are consistent with the real environment. Combined with UE4's blueprint system, dynamic interactive functions are given to each 3D model; Seamlessly integrate tunnel and building structures with terrain through custom modeling, and set dynamic weather and lighting conditions by integrating UE4's weather module and real-time lighting functions; Manage computing resources through the use of LOD systems, light maps, and instanced materials; Managing virtual scenes and setting data generation rules so that the rail transit virtual simulation system can provide different test scenes and output data; By using the UE4 Unreal Engine and visually modeling the three-dimensional model in a virtual environment, a rail transit virtual simulation system with weather parameters was constructed.

4. The method according to claim 1, characterized in that: Acquiring simulated image information sent by a simulated industrial camera and / or simulated sensor information sent by a simulated sensor, and performing visualization operations based on the rail transit virtual simulation system according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in rail transit scenes specifically includes: Acquire an industrial camera and / or sensor attached to a fixed position of the vehicle body with 6-DOF simulated by the Carla sensor simulation model, acquire simulated image information, and / or sensor high-simulation numbers, perform visualization operations based on the simulated image information and the sensor high-simulation numbers based on the rail transit virtual simulation system, display RGB, depth, semantic and event information, and key railway feature information, wherein the key railway feature information includes: trains, signals, sleepers and gantry, wherein the generation of sensor high-simulation numbers specifically includes: simulating sensor data through a sensor simulation simulator to generate sensor raw data, and based on the sensor correction model, according to actual needs, add noise, data attenuation and / or data distortion to the sensor raw data to obtain corrected sensor high-simulation data.

5. The method according to claim 1, characterized in that Creating a weather manager, and managing and controlling the weather parameters of the rail transit virtual simulation system through the weather manager specifically includes: A weather manager is created through Carla-UE4, corresponding weather parameters are input through the weather manager, weather conditions and model inputs are adjusted to manage and randomly generate weather conditions, and weather parameters of the rail transit virtual simulation system are managed and controlled.

6. A rail transit perception data simulation device based on virtual and real combination, characterized in that: include: A modeling module, used to determine the components of a rail transit scene, obtain real line data, perform modeling based on the components of the rail transit scene and the real line data, and establish a rail transit virtual simulation system with weather parameters; A weather module, used to create a weather manager, through which weather parameters of the rail transit virtual simulation system are managed and controlled; The visualization module is used to obtain simulated image information sent by simulated industrial cameras and / or simulated sensor information sent by simulated sensors. Based on the rail transit virtual simulation system, visualization operations are performed according to the simulated image information and the simulated sensor information to achieve visualization display of different industrial cameras and sensors in rail transit scenes.

7. The device according to claim 6, characterized in that The rail transit scene components specifically include: a basic element layer, a weather element layer, a lighting element layer, a road section element layer and a working condition element layer.

8. The device according to claim 6, characterized in that The modeling module is specifically used for: UE4 is used to model the rail transit scene components and the real line data. The height, undulation and details of the terrain in the rail transit scene are accurately restored by importing GIS data and using Landscape tools. The three-dimensional models of tracks, switches and sleepers are refined by using 3D modeling software. The Spline tool in UE4 is used for precise layout to ensure that the geometric parameters and operating characteristics of the rail system are consistent with the real environment. Combined with UE4's blueprint system, dynamic interactive functions are given to each 3D model; Seamlessly integrate tunnel and building structures with terrain through custom modeling, and set dynamic weather and lighting conditions by integrating UE4's weather module and real-time lighting functions; Manage computing resources through the use of LOD systems, light maps, and instanced materials; Managing virtual scenes and setting data generation rules so that the rail transit virtual simulation system can provide different test scenes and output data; Using the UE4 Unreal Engine, a three-dimensional model is visualized and designed in a virtual environment to build a rail transit virtual simulation system with weather parameters. The weather module is specifically used for: Create a weather manager through Carla-UE4, input corresponding weather parameters through the weather manager, adjust weather conditions and model inputs to manage and randomly generate weather conditions, and manage and control weather parameters of the rail transit virtual simulation system; The visualization module is specifically used for: Acquire an industrial camera and / or sensor attached to a fixed position of the vehicle body with 6-DOF simulated by the Carla sensor simulation model, acquire simulated image information, and / or simulated sensor information, and perform visualization operations based on the simulated image information and the simulated sensor information based on the rail transit virtual simulation system to display RGB, depth, semantic and event information, and key railway feature information, wherein the key railway feature information includes: trains, signals, sleepers and gantry, wherein the generation of sensor high-simulation numbers specifically includes: simulating sensor data through a sensor simulation simulator to generate sensor raw data, and based on the sensor correction model, according to actual needs, adding noise, data attenuation and / or data distortion to the sensor raw data to obtain corrected sensor high-simulation data.

9. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the rail transit perception data simulation method based on the combination of virtual and real as described in any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an implementation program for information transmission, and when the program is executed by the processor, the steps of the rail transit perception data simulation method based on the combination of virtual and real as described in any one of claims 1 to 5 are implemented.

Images