Simulation scene trueness evaluation method and device, equipment and storage medium

Abstract

The invention relates to the technical field of simulation scenes, and discloses a simulation scene trueness evaluation method, device and equipment and a storage medium, and the method comprises the steps: constructing a vehicle simulation scene, and carrying out the data collection of the vehicle simulation scene through a virtual sensor, and obtaining a simulation data set; acquiring operation data of the vehicle in a real scene through a real sensor to obtain a real data set; performing alignment processing on the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determining scene similarity based on the aligned simulation data set and the aligned real data set; and carrying out simulation scene authenticity evaluation according to the scene similarity to obtain an evaluation result. According to the invention, the evaluation accuracy of the trueness of the simulation scene can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of simulation scene technology, and in particular to a simulation scene reality evaluation method, device, equipment and storage medium. Background Art

[0002] In recent years, machine learning-based methods have been used to improve the realism and complexity of simulation scenarios. These methods are able to learn patterns from a large amount of historical data and generate new simulation scenarios. Although these scenarios are based on existing data sets, they introduce a certain degree of randomness, so they can generate scenarios that are different from historical data but still reasonable. However, existing methods usually rely on specific parameter data for calculations, which makes it difficult to ensure the authenticity and accuracy of the generated simulation scenarios. In addition, in some applications, such as autonomous vehicle testing, the authenticity and predictability of simulation scenarios are crucial. Although current technologies are capable of generating complex traffic scenarios, in some cases, due to the lack of full understanding and control of the generation process, the generated scenarios may deviate from the actual situation, and the authenticity of the simulation scenarios cannot be accurately evaluated, which in turn affects the effectiveness of the test.

[0003] Therefore, how to improve the accuracy of evaluating the realism of simulation scenes is a problem that needs to be solved urgently.

[0004] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of this application is to provide a method for evaluating the realism of a simulation scene, aiming to solve the technical problem of how to improve the accuracy of evaluating the realism of a simulation scene.

[0006] To achieve the above objectives, the present application proposes a method for evaluating the authenticity of a simulation scene, the method comprising:

[0007] Constructing a vehicle simulation scene, and collecting data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set;

[0008] Collect vehicle operation data in real scenarios through real sensors to obtain real data sets;

[0009] Aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set;

[0010] Determining scene similarity based on the aligned simulation data set and the aligned real data set;

[0011] The reality of the simulation scene is evaluated according to the scene similarity to obtain an evaluation result.

[0012] In one embodiment, constructing a vehicle simulation scenario includes:

[0013] Get scene data;

[0014] determining scene elements based on the scene data;

[0015] Based on the scene elements, environmental simulation and dynamic element addition are performed to generate a vehicle simulation scene.

[0016] In one embodiment, the aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set includes:

[0017] Acquire a first data acquisition parameter of the virtual sensor and a second data acquisition parameter of the real sensor;

[0018] The simulation data set and the real data set are aligned based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set.

[0019] In one embodiment, the aligning the simulation data set and the real data set based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set includes:

[0020] Determine a first acquisition frequency and a first timestamp according to the first data acquisition parameter, and determine a second acquisition frequency and a second timestamp according to the second data acquisition parameter;

[0021] When the first acquisition frequency and the second acquisition frequency are inconsistent, adjusting the data frequency in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set;

[0022] When the first timestamp and the second timestamp are inconsistent, data in the simulation data set and / or the real data set are reordered to obtain an aligned simulation data set and an aligned simulation data set.

[0023] In one embodiment, determining the scene similarity based on the aligned simulation data set and the aligned real data set includes:

[0024] Performing similarity calculation based on the aligned simulation data set and the aligned real data set to obtain a similarity set;

[0025] Performing data cleaning on the similarity set to obtain a cleaned similarity set;

[0026] The scene similarity is determined based on the cleaned similarity set.

[0027] In one embodiment, the similarity calculation is performed based on the aligned simulation data set and the aligned real data set to obtain a similarity set, including:

[0028] Calculating based on the simulated data in the aligned simulated data set and the corresponding real data in the aligned real data set to obtain the Euclidean distance between each simulated data and the corresponding real data;

[0029] Determine the similarity between each simulated data and the corresponding real data based on the Euclidean distance;

[0030] A similarity set is determined according to the similarities.

[0031] In one embodiment, the evaluating the reality of the simulation scene according to the scene similarity to obtain the evaluation result includes:

[0032] Comparing the scene similarity with a preset similarity threshold;

[0033] When the scene similarity is greater than or equal to a preset similarity threshold, determining that the evaluation result is that the simulation scene reality is qualified;

[0034] When the scene similarity is less than a preset similarity threshold, it is determined that the evaluation result is that the simulation scene reality is unqualified.

[0035] In addition, to achieve the above purpose, the present application also proposes a simulation scene reality evaluation device, the simulation scene reality evaluation device comprising:

[0036] Construction module, used to build vehicle simulation scenarios;

[0037] A collection module, used for collecting data of the vehicle simulation scene through a virtual sensor to obtain a simulation data set;

[0038] The acquisition module is also used to collect the operation data of the vehicle in the real scene through the real sensor to obtain the real data set;

[0039] A determination module, configured to determine scene similarity based on the simulation data set and the real data set;

[0040] The evaluation module is used to evaluate the reality of the simulation scene according to the scene similarity to obtain an evaluation result.

[0041] In addition, to achieve the above-mentioned purpose, the present application also proposes a simulation scene reality evaluation device, which includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and the computer program is configured to implement the steps of the simulation scene reality evaluation method as described above.

[0042] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the simulation scene reality evaluation method described above are implemented.

[0043] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the simulation scene reality evaluation method described above are implemented.

[0044] The present application provides a method for evaluating the authenticity of a simulation scene. The present application first constructs a vehicle simulation scene, and collects data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set; collects vehicle operation data in a real scene through a real sensor to obtain a real data set; aligns the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determines scene similarity based on the aligned simulation data set and the aligned real data set; and evaluates the authenticity of the simulation scene according to the scene similarity to obtain an evaluation result, which can effectively improve the accuracy of evaluating the authenticity of the simulation scene.

[0045] In summary, the present application collects data from vehicle simulation scenes and real scenes, obtains simulation data sets and real data sets and performs alignment processing, which can effectively improve the accuracy of scene similarity determination, and then accurately evaluate the reality of the simulation scene to obtain evaluation results, thus overcoming the technical defect of being unable to accurately evaluate the reality of the simulation scene, and can effectively improve the accuracy of the evaluation of the reality of the simulation scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0047] In order to more clearly illustrate the embodiments of the present application 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, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0048] Figure 1 A flowchart of the first embodiment of the method for evaluating the authenticity of a simulation scene provided in this application;

[0049] Figure 2 A flowchart of the second embodiment of the simulation scene reality evaluation method of the present application is provided;

[0050] Figure 3 A schematic diagram of the overall process of evaluating the authenticity of a simulation scene provided in Example 2 of the method for evaluating the authenticity of a simulation scene of the present application;

[0051] Figure 4 This is a schematic diagram of the module structure of the simulation scene reality evaluation device according to the embodiment of the present application;

[0052] Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the simulation scene reality evaluation method of the embodiment of the present application.

[0053] The purpose, features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0055] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0056] The main solution of the embodiment of the present application is: construct a vehicle simulation scene, and collect data for the vehicle simulation scene through virtual sensors to obtain a simulation data set; collect the operation data of the vehicle in the real scene through real sensors to obtain a real data set; align the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determine the scene similarity based on the aligned simulation data set and the aligned real data set; evaluate the authenticity of the simulation scene according to the scene similarity to obtain an evaluation result.

[0057] In recent years, machine learning-based methods have been used to improve the realism and complexity of simulation scenarios. These methods are able to learn patterns from a large amount of historical data and generate new simulation scenarios. Although these scenarios are based on existing data sets, they can generate scenarios that are different from historical data but still reasonable due to the introduction of a certain degree of randomness. However, existing methods usually rely on specific parameter data for calculation, which makes it difficult to ensure the realism and accuracy of the generated simulation scenarios. In addition, in some applications, such as autonomous vehicle testing, the realism and predictability of simulation scenarios are crucial. Although current technology can generate complex traffic scenarios, in some cases, due to the lack of full understanding and control of the generation process, the generated scenarios may deviate from the actual situation, and the realism of the simulation scenarios cannot be accurately evaluated, which in turn affects the effectiveness of the test. Therefore, how to improve the accuracy of the evaluation of the realism of simulation scenarios is a problem that needs to be solved urgently.

[0058] The present application collects data from vehicle simulation scenes and real scenes, obtains simulation data sets and real data sets, and performs alignment processing, which can effectively improve the accuracy of scene similarity determination, and then accurately evaluate the reality of the simulation scene to obtain evaluation results, thereby overcoming the technical defect of being unable to accurately evaluate the reality of the simulation scene, and can effectively improve the accuracy of the evaluation of the reality of the simulation scene.

[0059] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an electronic device capable of realizing the above functions, a simulation scene reality evaluation device, etc. The following takes the simulation scene reality evaluation device as an example to illustrate this embodiment and the following embodiments.

[0060] Based on this, the present application embodiment provides a simulation scene reality evaluation method, referring to Figure 1 , Figure 1 This is a flow chart of the first embodiment of the simulation scene reality evaluation method of the present application.

[0061] In this embodiment, the simulation scene reality evaluation method includes steps S10 to S50:

[0062] Step S10, constructing a vehicle simulation scene, and collecting data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set.

[0063] It should be noted that the vehicle simulation scene refers to the vehicle driving environment simulated by computer software or physical models, which is used to reproduce the traffic conditions, road conditions, weather changes and other factors in the real world, so as to evaluate vehicle performance, test autonomous driving algorithms, etc. In this embodiment, the construction of the vehicle simulation scene relies on high-precision map data, traffic flow models, physical engines, and vehicle dynamics models, etc., to ensure that the simulation environment can be as close to the real world as possible.

[0064] It is understandable that after the vehicle simulation scene is built, data is collected for the scene through virtual sensors. Virtual sensors are components that simulate the functions of real sensors in the simulation environment. They can sense and record various information in the simulation environment, such as vehicle position, speed, acceleration, distance and speed of surrounding objects, etc. Through the data collection of virtual sensors, a simulation data set containing rich information can be generated, which will serve as the basis for subsequent processing and analysis.

[0065] It is worth noting that the quality of the simulation data set directly affects the accuracy of the subsequent simulation scene reality evaluation. Therefore, during the data collection process, it is necessary to ensure the accuracy and reliability of the virtual sensor, and reasonably set the sampling frequency and data format so that the subsequent data processing and analysis can proceed smoothly.

[0066] In the specific implementation, after the simulation scene is constructed, the sensor configuration on the vehicle is further simulated. Through the interface provided by the simulation software, the virtual vehicle is equipped with a variety of high-precision sensors such as millimeter-wave radar, laser radar (LiDAR), global navigation satellite system (GNSS), etc. These sensors simulate the perception capabilities of the real world in the simulation environment, and receive and process information from the vehicle itself and its surrounding environment in real time, including key data such as vehicle position, speed, acceleration, and position and speed of road obstacles, i.e., simulation data.

[0067] It is worth noting that in order to ensure that the simulation data can be transmitted smoothly and used for subsequent analysis, a specific network port is set in the simulation software to output the data collected by the sensor. The data collected by the sensor exists in the form of raw data packets (RDB), which need to be parsed by unpacking tools to extract useful self-vehicle data (such as vehicle state parameters) and target vehicle / person data (such as position, speed, trajectory, etc.) and other simulation data. In addition, attention should be paid to the real-time and accuracy of simulation data transmission to ensure the consistency of time synchronization between simulation data and real-world data.

[0068] In a feasible implementation manner, "constructing a vehicle simulation scene" in step S10 specifically includes: acquiring scene data; determining scene elements based on the scene data; performing environmental simulation and adding dynamic elements based on the scene elements to generate a vehicle simulation scene.

[0069] It should be noted that scene data refers to the basic information used to build vehicle simulation scenes, including but not limited to map data, LiDAR data, image data, and sensor data. Map data can be used to obtain geographic information such as road layout and building outlines. Use LiDAR to scan the vehicle to obtain high-precision three-dimensional point cloud data of the road and surrounding environment. Obtain detailed information such as road markings and traffic signs through aerial images or street view images. Use various sensors (such as cameras, radars, ultrasonic sensors, etc.) to obtain real-time environmental data

[0070] It is understandable that after acquiring the scene data, the scene data is analyzed and processed to extract key scene elements. Scene elements refer to the basic units that constitute the simulation scene, such as roads, vehicles, pedestrians, buildings, trees, etc. Scene elements not only affect the appearance and layout of the simulation environment, but are also directly related to the accuracy and reliability of the simulation results.

[0071] It is worth mentioning that the work of environmental simulation and dynamic element addition is based on the extracted scene elements. Environmental simulation refers to the use of computer software or physical models to reproduce and present static elements in the simulation scene, including simulating lighting and weather. Among them, simulating lighting is to simulate the effects of sunlight in different time periods, taking into account the changes in the position and intensity of the sun; simulating weather is to simulate the effects of rain, snow, fog, sandstorms, etc. on vision and driving conditions. Adding dynamic elements refers to introducing elements with dynamic characteristics such as vehicles and pedestrians into the simulation scene to make the simulation environment more vivid and realistic, including traffic flow simulation and dynamic obstacle addition. Among them, traffic flow simulation refers to simulating real traffic flow and vehicle behavior, such as vehicle driving trajectory, pedestrian movement, etc.; dynamic obstacles refer to adding dynamic obstacles to the scene, such as temporary roadblocks, construction areas, etc., to increase the complexity of the scene.

[0072] In the specific implementation, in the professional simulation software platform, according to the real-world road layout, traffic signs, obstacle distribution and other details, through delicate manual operation or using the scene construction tool library, the elements of the real scene are reproduced one by one. This process not only requires high accuracy, but also needs to consider the impact of environmental factors such as lighting conditions and weather changes on scene perception to ensure the authenticity of the simulation environment. Through continuous adjustment and optimization, we strive to restore a virtual test environment that is highly similar to the real world in the simulation software, that is, the vehicle simulation environment.

[0073] Step S20, collecting the operation data of the vehicle in the real scene through real sensors to obtain a real data set.

[0074] It should be noted that real sensors are a key component in the vehicle testing and verification process. They can accurately capture the dynamic response and environmental parameters of the vehicle under various working conditions. Real sensors include but are not limited to speed sensors, acceleration sensors, gyroscopes, GPS positioning systems, and various environmental perception sensors (such as lidar, millimeter-wave radar, cameras, etc.). These sensors work together to comprehensively and meticulously record the driving status of the vehicle on the real road, including but not limited to speed, acceleration, direction, location information, and images, distance, speed, and other data of the surrounding environment, that is, real data.

[0075] It is understandable that in the process of collecting real data sets, the accuracy and reliability of sensors must be ensured. In order to obtain more comprehensive and rich data samples, test vehicles need to drive under different road conditions (such as highways, urban roads, rural roads, etc.), different traffic conditions (such as congestion, smooth traffic, complex intersections, etc.) and different weather and lighting conditions. Such data sets can not only reflect the performance of vehicles in different scenarios, but also provide simulation software with input data that is closer to the real world.

[0076] It is worth noting that in the process of real data collection, attention should also be paid to the real-time and synchronization of real data. That is, the data between various sensors must be highly consistent to accurately reflect the status of the vehicle at the same time, which usually requires the use of high-precision clock synchronization technology or data fusion algorithms to achieve this.

[0077] Step S30, aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set.

[0078] It should be noted that the purpose of the alignment process is to ensure the consistency of the simulated dataset and the real dataset in terms of time series, spatial position and data dimension, so that effective comparative analysis can be carried out later. Only when the two datasets are aligned at multiple levels can the accuracy and reliability of the simulation model be accurately evaluated.

[0079] It is understandable that in order to achieve efficient alignment processing, a variety of technical means can be used. First, the time stamp information can be used to align the two data sets in time to ensure that each frame of simulation data is closely connected with the corresponding real data in time. The driving trajectory of the simulated vehicle can also be aligned in space with the driving trajectory of the real vehicle through the GPS positioning system or map matching technology to eliminate the error caused by position deviation. In addition, for sensor data, such as speed, acceleration, direction, etc., data fusion algorithms or interpolation methods can be used to ensure that these data can also maintain consistency in numerical values, and this embodiment does not make specific restrictions on this.

[0080] It is worth noting that during the alignment process, attention should be paid to the integrity and continuity of the data. Due to the possible differences between the simulation environment and the real environment, some data in the simulation may not be completely matched to the real situation. To solve this problem, data enhancement technology can be used to generate additional simulation data to fill these blank areas. At the same time, similar methods can be used to recover the missing real data due to sensor failure or data acquisition problems.

[0081] In a feasible implementation, step S30 may include: obtaining a first data acquisition parameter of the virtual sensor and a second data acquisition parameter of the real sensor; aligning the simulation data set and the real data set based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set.

[0082] It should be noted that the data collection parameters refer to specific indicators and settings for guiding the sensor on how to collect and process data, such as the collection frequency and timestamp, which are not specifically limited in this embodiment. The first data collection parameters include a first collection frequency and a first timestamp, and the second data collection parameters include a second collection frequency and a second timestamp.

[0083] It is understandable that the first data acquisition parameter of the virtual sensor should simulate the characteristics of the real sensor as much as possible, so that the simulation data can truly reflect the physical quantity in the actual driving environment. For example, if the real sensor collects speed data at a frequency of 100 times per second, the data acquisition frequency of the virtual sensor should also be set to 100 times per second to maintain the consistency of the data in the time series. In addition, it is also necessary to ensure that the timestamp information of the virtual sensor and the real sensor can correspond to each other, so that the simulation data can be accurately matched with the real data in subsequent processing.

[0084] In the specific implementation, on the basis of data acquisition parameter alignment, advanced data processing technology can be further used to optimize the alignment effect. For example, the Dynamic Time Warping (DTW) algorithm can be used to process small offsets or deformations in time series data, so that the simulation data and real data are more closely matched in time. The DTW algorithm can automatically adjust the time axis to minimize the distance between the two time series by calculating the optimal matching path between them, thereby improving the accuracy of the alignment.

[0085] In a feasible implementation manner, the alignment processing of the simulation data set and the real data set based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set includes: determining a first acquisition frequency and a first timestamp according to the first data acquisition parameter, and determining a second acquisition frequency and a second timestamp according to the second data acquisition parameter; when the first acquisition frequency and the second acquisition frequency are inconsistent, adjusting the data frequency in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set; when the first timestamp and the second timestamp are inconsistent, reordering the data in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set.

[0086] It should be noted that the acquisition frequency describes the speed at which the sensor collects data, that is, the number of data points collected per unit time. For example, if the sensor collects data at a frequency of 100 times per second, its data acquisition frequency is 100Hz. In the scenario where the autonomous driving simulation is compared with the real environment, it is very important to ensure the consistency of the acquisition frequency between the simulation data set and the real data set, which not only maintains the continuity of the data in the time series, but also ensures the accuracy of the subsequent analysis results.

[0087] It is understandable that when it is found that the first acquisition frequency (i.e., the acquisition frequency of the virtual sensor) is inconsistent with the second acquisition frequency (i.e., the acquisition frequency of the real sensor), appropriate adjustment measures need to be taken. If the acquisition frequency of the simulated data set is higher than that of the real data set, the sample points in the simulated data can be reduced by downsampling technology to match the acquisition frequency of the real data. On the contrary, if the acquisition frequency of the simulated data set is lower, it may be necessary to increase the sample points in the simulated data by interpolation methods (such as linear interpolation, polynomial interpolation, etc.) to improve its time resolution so as to be consistent with the real data set.

[0088] It is worth mentioning that the alignment of timestamps is equally important. Timestamps are the time mark of each sample point in the data set, which records the exact moment when the data was collected. In autonomous driving simulation, since there may be slight delays or asynchrony between the simulation environment and the real environment, the alignment of timestamps is a key step to ensure that the data accurately corresponds to the timeline. When it is found that the first timestamp is inconsistent with the second timestamp, the data needs to be reordered or the timeline adjusted to ensure that the simulation data set and the real data set can closely match in time.

[0089] To further optimize the timestamp alignment process, the timestamps in the two data sets can be preprocessed, including removing noisy timestamps (such as abnormal timestamps caused by system errors) and filling in missing timestamps (through interpolation or other prediction methods). Then, a high-precision time synchronization algorithm, such as the Network Time Protocol (NTP) or the Precision Time Protocol (PTP), is used to calibrate the time difference between the simulation system and the real system to ensure that the simulation data is highly consistent with the real-world time when it is generated. In the data alignment stage, by comparing the timestamps in the simulation data set with the timestamps in the real data set, the simulation data is adjusted accordingly, such as time shifting or time scaling, to eliminate small deviations in time.

[0090] In the specific implementation, in order to effectively compare the similarity between real data and simulated data, it is necessary to ensure that the two are comparable in key scoring items, that is, the basic information such as the frequency and timestamp of the sensor data needs to be aligned. If the real sensor and the ideal sensor in the simulation are consistent in data collection frequency, and the data within the 5 to 10 second time window before the corresponding event in the two scenes are selected for alignment, it lays the foundation for the subsequent similarity calculation. If the common scoring items cannot be identified, it is considered that the construction of the simulation scene has failed, and there is no need to perform similarity calculation.

[0091] Step S40: determining scene similarity based on the aligned simulation data set and the aligned real data set.

[0092] It should be noted that scene similarity refers to the degree of similarity between the simulation data set and the real data set in a specific scenario, and is a key indicator for measuring the effectiveness of vehicle simulation scenarios.

[0093] It is understandable that in order to accurately calculate the scene similarity, it is necessary to first define a series of scoring criteria or metrics that should be able to fully reflect the differences and similarities of the data sets in key aspects. For example, the vehicle's driving trajectory, speed change, acceleration, steering angle and other kinematic parameters, as well as surrounding environment information such as obstacle locations, road signs, traffic lights, etc. can be considered. The degree of difference in these parameters and information will directly affect the calculation results of the scene similarity.

[0094] It is worth noting that, in the calculation process, a variety of mathematical methods and algorithms can be used to quantify these differences. For example, distance measurement methods such as Euclidean distance, Manhattan distance or cosine similarity can be used to calculate the similarity between two data sets in a specific dimension. This embodiment does not specifically limit this, and this embodiment uses Euclidean distance as an example for explanation.

[0095] Step S50, evaluating the reality of the simulation scene according to the scene similarity to obtain an evaluation result.

[0096] It should be noted that the evaluation of the reality of the simulation scene is to determine whether the reality of the simulation scene is qualified. According to the evaluation results, the simulation scene can be further optimized and adjusted. If the evaluation results show that the reality of the simulation scene is unqualified, that is, the reality of the simulation scene is low, that is, there is a large difference from the actual situation, then the key parameters in the simulation scene should be reviewed and adjusted to ensure that they can more accurately reflect the physical laws and behavioral characteristics of the real world, and deeply integrate the simulation scene with the actual data to correct the deviations and errors in the simulation process.

[0097] In a feasible implementation, step S50 may include: comparing the scene similarity with a preset similarity threshold; when the scene similarity is greater than or equal to the preset similarity threshold, determining that the evaluation result is that the simulation scene realism is qualified; when the scene similarity is less than the preset similarity threshold, determining that the evaluation result is that the simulation scene realism is unqualified.

[0098] It should be noted that the preset similarity threshold may be 90% or other values, and this embodiment does not impose any specific limitation on this. This embodiment is described by taking 90% as an example.

[0099] It is understandable that a reasonable threshold (such as 90%) is set to determine whether the similarity between the simulation scene and the real scene meets the requirements. When the scene similarity reaches or exceeds the threshold, the authenticity of the simulation scene is considered to be qualified and has a high similarity with the original real scene.

[0100] The present embodiment provides a method for evaluating the authenticity of a simulation scene. The present embodiment first constructs a vehicle simulation scene, and collects data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set; collects the operation data of the vehicle in the real scene through a real sensor to obtain a real data set; aligns the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determines the scene similarity based on the aligned simulation data set and the aligned real data set; and evaluates the authenticity of the simulation scene according to the scene similarity to obtain an evaluation result, which can effectively improve the accuracy of evaluating the authenticity of the simulation scene.

[0101] In summary, this embodiment collects data from vehicle simulation scenes and real scenes to obtain simulation data sets and real data sets and aligns them, which can effectively improve the accuracy of scene similarity determination, and then accurately evaluate the reality of the simulation scene to obtain an evaluation result, thereby overcoming the technical defect of being unable to accurately evaluate the reality of the simulation scene, and can effectively improve the accuracy of the evaluation of the reality of the simulation scene.

[0102] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can refer to the above introduction, and will not be repeated later. Figure 2 , the step S40 further includes steps S401-S403:

[0103] Step S401 , performing similarity calculation based on the aligned simulation data set and the aligned real data set to obtain a similarity set.

[0104] It should be noted that the similarity set refers to the set of similarity values ​​between each frame of data in the aligned simulation data set and the corresponding data in the real data set. These similarity values ​​reflect the degree of matching between the simulation scene and the real scene in different dimensions or features.

[0105] It is understandable that similarity can be calculated by Euclidean distance, Manhattan distance, cosine similarity or other measurement methods, and this embodiment does not specifically limit this. This embodiment is described by taking Euclidean distance as an example. Euclidean distance is the real distance between two points in multidimensional space, which is the straight-line distance. In similarity calculation, a shorter Euclidean distance indicates a higher similarity, while a longer distance indicates a lower similarity. Therefore, in this embodiment, the Euclidean distance is used to calculate the similarity between the aligned simulation data set and the real data set, and a similarity set is generated.

[0106] In a feasible implementation, step S401 may include: calculating based on the simulated data in the aligned simulation data set and the corresponding real data in the aligned real data set to obtain the Euclidean distance between each simulated data and the corresponding real data; determining the similarity between each simulated data and the corresponding real data based on the Euclidean distance; and determining a similarity set based on the similarity.

[0107] It should be noted that after the common scoring items are established, the Euclidean distance calculation formula is used as the main tool for similarity measurement. This formula measures the similarity between two data sets by calculating the difference in the corresponding sensor parameters in each frame of data. The Euclidean distance calculation formula is as follows:

[0108]

[0109] Where d(x,y) is the Euclidean distance between the simulated data x and the real data y.

[0110] It can be understood that by calculating the Euclidean distance for each frame of data one by one and taking the Euclidean distance as the similarity, a collection of similarities of all frame data (d1, d2, ..., dn ), the similarity collection provides an intuitive understanding of the overall scene similarity.

[0111] Step S402: performing data cleaning on the similarity set to obtain a cleaned similarity set.

[0112] It should be noted that the purpose of data cleaning is to remove or correct errors, outliers, duplicates, etc. in the data to improve the accuracy and reliability of subsequent data analysis.

[0113] It is understandable that in the process of data cleaning, outliers in the similarity set are identified and processed. Outliers may be caused by errors in the data alignment process, sensor failures, or special conditions during data collection. These outliers may distort the overall distribution of similarities. Therefore, statistical methods (such as box plots, Z-scores, etc.) are needed to identify and remove them. For the processing of outliers, you can choose to directly delete or replace them with reasonable estimates, depending on the characteristics of the data and the needs of the analysis. Check and remove duplicates in the similarity set. Although the similarity of each frame of data should be unique in theory, in actual operation, duplicates may appear due to errors in data processing or storage. These duplicates not only increase the burden of data processing, but may also mislead subsequent analysis results. Therefore, it is necessary to ensure the uniqueness of the similarity set through comparison and deduplication operations.

[0114] In addition, the similarity set needs to be normalized. Normalization is a process of scaling data so that it falls into a small specific interval (such as 0 to 1). In a similarity set, since the similarities of different frame data may have different dimensions and magnitudes, direct comparison may not be fair. Through normalization, this difference can be eliminated, making the comparison between similarities more objective and accurate.

[0115] Finally, the quality of the cleaned similarity set can be evaluated by calculating the statistical indicators of the set (such as mean, standard deviation, median, etc.) to understand the overall distribution and dispersion of the similarity. The cleaned similarity set can also be compared with the original data to evaluate the effect and accuracy of data cleaning.

[0116] In the specific implementation, after obtaining the similarity set, the data is further cleaned and optimized. First, identify and process abnormal data points in the set, such as similarity values ​​greater than 1 or less than 0 (which is unreasonable under normal circumstances), and reduce their impact on the overall results by processing missing values ​​or directly deleting these abnormal values. Then, the 3σ principle in statistics is used to filter and correct the data in the set. For data points beyond the 3σ range, the similarity value of the previous frame is selected to replace them to ensure the continuity and stability of the data.

[0117] Step S403: determining scene similarity based on the cleaned similarity set.

[0118] It should be noted that after data cleaning and optimization, the similarity of the entire scene is recalculated based on the cleaned data set, aiming to evaluate the similarity between different scenes through precise metrics. When determining scene similarity, a series of complex algorithms and models are usually used. These tools can capture and quantify the subtle differences and common features between scenes.

[0119] In the specific implementation, the cleaned similarity set is deeply analyzed to identify the key similarities and differences between scenes, which involves a detailed interpretation of the similarity values ​​and a comprehensive consideration of the scene content, structure, dynamic changes, etc. Through these analyses, a multi-dimensional similarity evaluation framework can be constructed, which can fully reflect the similarity between scenes. Next, advanced technologies such as machine learning or deep learning are used to model and predict the similarity set. These models can automatically learn the complex relationships between scenes and predict the similarity between new scenes and known scenes based on these relationships. In this way, the accuracy and efficiency of similarity evaluation can be greatly improved.

[0120] like Figure 3 As shown, Figure 3 This is a schematic diagram of the overall process of evaluating the authenticity of a simulation scene. The overall process of evaluating the authenticity of a simulation scene includes obtaining real scene data; setting up the scene and installing perfect sensors to obtain a simulation scene and collect simulation scene data; aligning the simulation data with the real data, and then calculating the similarity of each frame of data to obtain a similarity data set; performing data cleaning on the similarity data set, and determining the final similarity of the scene based on the cleaned similarity data set.

[0121] In this embodiment, by performing similarity calculation based on the aligned simulation data set and the aligned real data set, the data accuracy and completeness of the similarity set are effectively improved, the similarity set is data cleansed, and the scene similarity is determined based on the cleaned similarity set, which effectively improves the scene similarity accuracy, thereby accurately evaluating the realism of the simulation scene.

[0122] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the simulation scene reality evaluation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0123] This application also provides a simulation scene reality evaluation device, please refer to Figure 3 , the simulation scene reality evaluation device comprises:

[0124] The construction module 10 is used to construct a vehicle simulation scene and collect data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set.

[0125] The acquisition module 20 is used to collect the operation data of the vehicle in the real scene through real sensors to obtain a real data set.

[0126] The alignment module 30 is used to align the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set.

[0127] The determination module 40 is used to determine the scene similarity based on the aligned simulation data set and the aligned real data set.

[0128] The evaluation module 50 is used to evaluate the reality of the simulation scene according to the scene similarity to obtain an evaluation result.

[0129] The present embodiment provides a simulation scene reality evaluation device, which first constructs a vehicle simulation scene, and collects data on the vehicle simulation scene through virtual sensors to obtain a simulation data set; collects vehicle operation data in a real scene through real sensors to obtain a real data set; aligns the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determines scene similarity based on the aligned simulation data set and the aligned real data set; and evaluates the simulation scene reality according to the scene similarity to obtain an evaluation result, which can effectively improve the accuracy of evaluating the simulation scene reality.

[0130] In summary, this embodiment collects data from vehicle simulation scenes and real scenes to obtain simulation data sets and real data sets and aligns them, which can effectively improve the accuracy of scene similarity determination, and then accurately evaluate the reality of the simulation scene to obtain an evaluation result, thereby overcoming the technical defect of being unable to accurately evaluate the reality of the simulation scene, and can effectively improve the accuracy of the evaluation of the reality of the simulation scene.

[0131] Optionally, the construction module 10 is further used to obtain scene data; determine scene elements based on the scene data; perform environmental simulation and add dynamic elements based on the scene elements to generate a vehicle simulation scene.

[0132] Optionally, the alignment module 30 is also used to obtain a first data acquisition parameter of the virtual sensor and a second data acquisition parameter of the real sensor; based on the first data acquisition parameter and the second data acquisition parameter, the simulation data set and the real data set are aligned to obtain an aligned simulation data set and an aligned simulation data set.

[0133] Optionally, the alignment module 30 is also used to determine a first acquisition frequency and a first timestamp based on the first data acquisition parameter, and to determine a second acquisition frequency and a second timestamp based on the second data acquisition parameter; when the first acquisition frequency and the second acquisition frequency are inconsistent, adjust the data frequency in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set; when the first timestamp and the second timestamp are inconsistent, reorder the data in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set.

[0134] Optionally, the determination module 40 is further used to perform similarity calculation based on the aligned simulation data set and the aligned real data set to obtain a similarity set; perform data cleaning on the similarity set to obtain a cleaned similarity set; and determine scene similarity based on the cleaned similarity set.

[0135] Optionally, the determination module 40 is also used to calculate based on the simulated data in the aligned simulation data set and the corresponding real data in the aligned real data set to obtain the Euclidean distance between each simulated data and the corresponding real data; determine the similarity between each simulated data and the corresponding real data based on the Euclidean distance; and determine a similarity set based on the similarity.

[0136] Optionally, the evaluation module 50 is further used to compare the scene similarity with a preset similarity threshold; when the scene similarity is greater than or equal to the preset similarity threshold, the evaluation result is determined to be that the simulation scene reality is qualified; when the scene similarity is less than the preset similarity threshold, the evaluation result is determined to be that the simulation scene reality is unqualified.

[0137] The simulation scene reality evaluation device provided by the present application adopts the simulation scene reality evaluation method in the above embodiment, which can solve the technical problem of how to improve the evaluation accuracy of the simulation scene reality. Compared with the prior art, the beneficial effects of the simulation scene reality evaluation device provided by the present application are the same as the beneficial effects of the simulation scene reality evaluation method provided by the above embodiment, and the other technical features in the simulation scene reality evaluation device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0138] The present application provides a simulation scene reality evaluation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the simulation scene reality evaluation method in the above-mentioned embodiment one.

[0139] Reference below Figure 4 , which shows a schematic diagram of the structure of a simulation scene reality evaluation device suitable for implementing the embodiment of the present application. The simulation scene reality evaluation device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The simulation scene reality evaluation device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0140] like Figure 4As shown, the simulation scene reality evaluation device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the simulation scene reality evaluation device are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 can allow the simulation scene reality evaluation device to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows a simulation scene reality evaluation device with various systems, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0141] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0142] The simulation scene reality evaluation device provided by the present application adopts the simulation scene reality evaluation method in the above embodiment, which can solve the technical problem of how to improve the evaluation accuracy of the simulation scene reality. Compared with the prior art, the beneficial effects of the simulation scene reality evaluation device provided by the present application are the same as the beneficial effects of the simulation scene reality evaluation method provided by the above embodiment, and the other technical features in the simulation scene reality evaluation device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0143] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0144] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0145] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, and the computer-readable program instructions are used to execute the simulation scene reality evaluation method in the above-mentioned embodiment.

[0146] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination of the above.

[0147] The computer-readable storage medium may be included in the simulation scene reality evaluation device; or may exist independently without being assembled into the simulation scene reality evaluation device.

[0148] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by the simulation scene reality evaluation device, the simulation scene reality evaluation device is enabled to: construct a vehicle simulation scene, and collect data for the vehicle simulation scene through virtual sensors to obtain a simulation data set; collect the operation data of the vehicle in the real scene through real sensors to obtain a real data set; align the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; determine the scene similarity based on the aligned simulation data set and the aligned real data set; and evaluate the simulation scene reality according to the scene similarity to obtain an evaluation result.

[0149] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0150] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0151] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0152] The readable storage medium provided by the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned simulation scene authenticity evaluation method, and can solve the technical problem of how to improve the accuracy of the evaluation of the simulation scene authenticity. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as the beneficial effects of the simulation scene authenticity evaluation method provided by the above-mentioned embodiment, and will not be repeated here.

[0153] The present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned simulation scene reality evaluation method when executed by a processor.

[0154] The computer program product provided by the present application can solve the technical problem of how to improve the accuracy of the evaluation of the realism of the simulation scene. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the simulation scene realism evaluation method provided by the above embodiment, which will not be repeated here.

[0155] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for evaluating the authenticity of a simulation scene, characterized in that: The method comprises: Constructing a vehicle simulation scene, and collecting data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set; Collect vehicle operation data in real scenarios through real sensors to obtain real data sets; Aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; Determining scene similarity based on the aligned simulation data set and the aligned real data set; The reality of the simulation scene is evaluated according to the scene similarity to obtain an evaluation result.

2. The method according to claim 1, characterized in that The constructing of the vehicle simulation scene includes: Get scene data; determining scene elements based on the scene data; Based on the scene elements, environmental simulation and dynamic element addition are performed to generate a vehicle simulation scene.

3. The method according to claim 1, characterized in that The aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set includes: Acquire a first data acquisition parameter of the virtual sensor and a second data acquisition parameter of the real sensor; The simulation data set and the real data set are aligned based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set.

4. The method according to claim 3, characterized in that The aligning the simulation data set and the real data set based on the first data acquisition parameter and the second data acquisition parameter to obtain an aligned simulation data set and an aligned simulation data set includes: Determine a first acquisition frequency and a first timestamp according to the first data acquisition parameter, and determine a second acquisition frequency and a second timestamp according to the second data acquisition parameter; When the first acquisition frequency and the second acquisition frequency are inconsistent, adjusting the data frequency in the simulation data set and / or the real data set to obtain an aligned simulation data set and an aligned simulation data set; When the first timestamp and the second timestamp are inconsistent, data in the simulation data set and / or the real data set are reordered to obtain an aligned simulation data set and an aligned simulation data set.

5. The method according to claim 1, characterized in that The determining of scene similarity based on the aligned simulation data set and the aligned real data set includes: Performing similarity calculation based on the aligned simulation data set and the aligned real data set to obtain a similarity set; Performing data cleaning on the similarity set to obtain a cleaned similarity set; The scene similarity is determined based on the cleaned similarity set.

6. The method according to claim 5, characterized in that The similarity calculation is performed based on the aligned simulation data set and the aligned real data set to obtain a similarity set, including: Calculating based on the simulated data in the aligned simulated data set and the corresponding real data in the aligned real data set to obtain the Euclidean distance between each simulated data and the corresponding real data; Determine the similarity between each simulated data and the corresponding real data based on the Euclidean distance; A similarity set is determined according to the similarities.

7. The method according to claim 1, characterized in that The evaluating the reality of the simulation scene according to the scene similarity to obtain the evaluation result includes: Comparing the scene similarity with a preset similarity threshold; When the scene similarity is greater than or equal to a preset similarity threshold, determining that the evaluation result is that the simulation scene reality is qualified; When the scene similarity is less than a preset similarity threshold, it is determined that the evaluation result is that the simulation scene reality is unqualified.

8. A simulation scene reality evaluation device, characterized in that: The simulation scene reality evaluation device comprises: A construction module is used to construct a vehicle simulation scene and collect data on the vehicle simulation scene through a virtual sensor to obtain a simulation data set; The acquisition module is used to collect the operation data of the vehicle in the real scene through real sensors to obtain a real data set; An alignment module, used for aligning the simulation data set and the real data set to obtain an aligned simulation data set and an aligned simulation data set; A determination module, configured to determine scene similarity based on the aligned simulation data set and the aligned real data set; The evaluation module is used to evaluate the reality of the simulation scene according to the scene similarity to obtain an evaluation result.

9. A simulation scene reality evaluation device, characterized in that: The simulation scene reality evaluation device comprises: a memory, a processor and a simulation scene reality evaluation program stored in the memory and executable on the processor, wherein the simulation scene reality evaluation program is configured to implement the simulation scene reality evaluation method according to any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a simulation scene reality evaluation program, which, when executed by a processor, implements a simulation scene reality evaluation method according to any one of claims 1 to 7.

Images