A method for generating scene library data

Abstract

The present invention relates to the technical field of autonomous driving simulation scene construction, and in particular to a scene library data generation method, which comprises the following steps: using ordinary GPS interval setting time to obtain positioning information of a collection vehicle when it is driving; obtaining distance data of environmental targets in front of the collection vehicle relative to the collection vehicle; establishing a global coordinate system, fitting the driving trajectory of the collection vehicle according to GPS data, and fitting the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle. This solution can solve the problem of using expensive high-line laser radars in the prior art, which increases costs due to high collection requirements and powerful performance requirements of supporting collection equipment and storage devices.

Description

Technical Field

[0001] The present invention relates to the technical field of autonomous driving simulation scene construction, and in particular to a scene library data generation method. Background Art

[0002] Driving scenario data is a basic data resource for the development and testing of intelligent connected vehicles. It is an important "case library" and "exercise book" for evaluating the functional safety of intelligent connected vehicles, and a key data basis for redefining the level of intelligent vehicles. Driving scenario data is a basic data resource for the development and testing of intelligent connected vehicles. It is an important "case library" and "exercise book" for evaluating the functional safety of intelligent connected vehicles, and a key data basis for redefining the level of intelligent vehicles. Driving scenario test cases are mainly reproduced through virtual simulation environments and tool chains. Therefore, building a virtual scenario database is a key bridge connecting scenario data and scenario applications. The virtual scenario database is characterized by infinity, scalability, batching, and automation.

[0003] The main contents of the simulation scene library construction are: based on the actual road scene, the corresponding simulation scene is constructed in the simulation software. The main process includes: real vehicle data collection, real vehicle data processing, data analysis and extraction, scene parameterization definition, and scene construction in the simulation scene according to the parameterization definition. Finally, in the constructed simulation scene, the real-time position relationship of the target object relative to the vehicle should be highly consistent with the actual road, so as to support sensor recognition testing and assisted driving / automatic driving function testing on the simulation bench.

[0004] General solution of existing technology: The existing scene library real-vehicle data acquisition system generally includes high-precision lidar (number of lines ≥ 64), inertial navigation GPS / differential GPS (accuracy at the cm level), smart cameras, millimeter-wave radar, driving recorders and other equipment to achieve high-precision and high-frequency data acquisition.

[0005] Existing technologies use high-beam laser radar and inertial navigation GPS / differential GPS. Although they can obtain relatively accurate data, high-precision and high-frequency data collection leads to excessive data volume, high requirements for the performance of acquisition equipment and storage devices, resulting in high costs, and a huge workload for subsequent data processing, resulting in high workload and low efficiency. In addition, the acquisition cost is high: the use of expensive high-line laser radars requires high performance of supporting acquisition equipment and storage devices, which increases costs. Summary of the invention

[0006] In view of the defects existing in the prior art, the purpose of the present invention is to provide a scene library data generation method, which can solve the problem of the prior art using expensive high-line-count laser radar, which has high acquisition requirements and requires powerful supporting acquisition equipment and storage devices, resulting in increased costs.

[0007] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0008] The present invention provides a method for generating scene library data, comprising the following steps:

[0009] Use ordinary GPS interval setting time to obtain the location information of the collection vehicle when it is driving;

[0010] Obtaining distance data of environmental targets in front of the collection vehicle relative to the collection vehicle;

[0011] A global coordinate system is established, and the driving trajectory of the collection vehicle is fitted according to the GPS data. The trajectory of the environmental target is fitted according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle.

[0012] In some optional solutions, the driving trajectory of the collection vehicle is fitted according to the GPS data, including:

[0013] Establish a global coordinate system and convert the positioning information obtained by the collection vehicle into global coordinates;

[0014] The global coordinates of all positioning information collected by the vehicle during driving are divided into segments, and the trajectory curve is fitted;

[0015] The segmented fitting trajectory curve is smoothed to obtain the driving trajectory of the collection vehicle in the global coordinate system.

[0016] In some optional solutions, distance data of the environmental target in front of the collection vehicle relative to the collection vehicle is obtained, including lateral distance, longitudinal distance and angle information of the environmental target relative to the collection vehicle.

[0017] In some optional solutions, the trajectory of the environmental target is fitted according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle, including:

[0018] On the driving track of the collection vehicle, the lateral distance and longitudinal distance of each sampling point of the environmental target relative to the driving track of the collection vehicle are superimposed to obtain the global coordinate discrete point of the environmental target;

[0019] The global coordinate discrete points of the environmental target are fitted with the global coordinate discrete points of the environmental target through piecewise polynomials to obtain the piecewise trajectory of the global coordinate discrete points of the environmental target;

[0020] The uneven parts of the segmented splicing of the segmented trajectory of the global coordinate discrete points of the environmental target are smoothed to obtain the trajectory of the environmental target in the global coordinate system.

[0021] In some optional solutions, obtaining distance data of an environmental target in front of a collection vehicle relative to the collection vehicle includes the following steps:

[0022] Two smart cameras and two millimeter-wave radars are used to collect data on the lateral collection distance and longitudinal collection distance of environmental targets relative to the collection vehicle, and an ordinary camera is used to obtain video information of environmental targets in front of the collection vehicle;

[0023] The validity of the lateral and longitudinal collection distances of the environmental targets relative to the collection vehicle collected by the two smart cameras and the two millimeter-wave radars is judged. The video information of the environmental targets in front of the collection vehicle is obtained by combining with ordinary cameras, and invalid data is filtered out and data processing is performed to obtain the lateral and longitudinal distances of the environmental targets relative to the collection vehicle.

[0024] In some optional solutions, the data collected by two smart cameras and two millimeter-wave radars about the lateral collection distance and longitudinal collection distance of the environmental target relative to the collection vehicle are judged for validity, and the video information of the environmental target in front of the collection vehicle is obtained by combining with the ordinary camera, and the invalid data is screened out and the data is processed to obtain the lateral distance and longitudinal distance of the environmental target relative to the collection vehicle, including:

[0025] When determining the longitudinal distance of an environmental target relative to the collection vehicle:

[0026] First, the longitudinal distance data of the environmental targets collected by the two millimeter-wave radars relative to the collection vehicle are compared.

[0027] When the difference between the longitudinal data collected by the two millimeter-wave radars is within the first set range, the average of the longitudinal data collected by the two millimeter-wave radars is used as the longitudinal distance of the environmental target relative to the collection vehicle;

[0028] When the difference in the longitudinal collection distance data collected by the two millimeter-wave radars is not within the first set range, the mean longitudinal collection distance data of the environmental target collected by the two smart cameras relative to the collection vehicle is compared with the mean longitudinal collection distance data of the environmental target collected by the two millimeter-wave radars relative to the collection vehicle, and the one with the smaller error with the mean longitudinal collection distance data of the two smart cameras is taken as the final longitudinal distance of the environmental target relative to the collection vehicle;

[0029] When determining the lateral distance of an environmental target relative to the collection vehicle:

[0030] First, compare the lateral collection distance data of the environmental targets of the two smart cameras relative to the collection vehicle.

[0031] When the difference between the lateral data collected by the two smart cameras is within the second set range, the average of the lateral data collected by the two smart cameras is used as the lateral distance of the environmental target relative to the collection vehicle;

[0032] When the difference in the lateral collection distance data collected by the two intelligent cameras is not within the second set range, an ordinary camera is used to obtain video information of the environmental target in front of the collection vehicle for manual re-judgment to determine the lateral distance of the environmental target relative to the collection vehicle.

[0033] In some optional schemes, when two millimeter-wave radars are used to collect data on the lateral collection distance and the longitudinal collection distance of the environmental target relative to the collection vehicle, the RCS of the environmental vehicles and pedestrians in the environmental target are also collected.

[0034] In some optional schemes, when two smart cameras are used to collect data on the lateral collection distance and longitudinal collection distance of the environmental target relative to the collection vehicle, contour information of environmental vehicles, pedestrians and lane lines in the environmental target is also collected and type classification is performed.

[0035] In some optional solutions, the orientation angles of the environmental vehicles and pedestrians are determined according to the trajectories of the environmental vehicles and pedestrians in the trajectory of the fitted environmental target.

[0036] In some optional schemes, when obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the laser radar is also used to collect the relative distance data between the environmental target and the collection vehicle. According to the relative distance data between the environmental target and the collection vehicle collected by the laser radar, the absolute distance between the driving trajectory of the collection vehicle and the trajectory of the environmental target in the global coordinate system is checked. Compared with the prior art, the advantages of the present invention are: the positioning information of the collection vehicle when driving is obtained by using ordinary satellite positioning GPS (meter level) instead of the expensive inertial navigation GPS or differential GPS in the existing scheme, and when obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the frequency of acquisition is close to the positioning time interval of ordinary GPS, which can reduce the performance requirements of the collection equipment and storage equipment, thereby forming a low-cost solution. The trajectory of the collection vehicle in the global coordinate system is obtained by fitting, and the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle is obtained, combined with the trajectory of the collection vehicle in the global coordinate system, the data of the environmental target in the global coordinate system is obtained, and then fitting is performed to obtain the trajectory of the collection vehicle and the environmental target in the global coordinate system, and the data required for the simulation database can also be generated more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0038] Figure 1A schematic diagram of a flow chart of a method for generating scene library data in an embodiment of the present invention;

[0039] Figure 2 A schematic diagram of fitting the trajectory of a collection vehicle in an embodiment of the present invention;

[0040] Figure 3 Schematic diagram of fitting the trajectory of an environmental target in an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0042] The embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0043] like Figure 1 As shown, the present invention provides a method for generating scene library data, comprising the following steps:

[0044] S1: Use ordinary GPS interval setting time to obtain the positioning information of the collection vehicle while it is driving.

[0045] The location information of the collection vehicle when driving includes the coordinates in the global positioning system and the vehicle speed information. The coordinates in the global positioning system are obtained through ordinary GPS, and the vehicle speed information can be obtained through ordinary GPS or the bus of the collection vehicle.

[0046] S2: Obtain the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle.

[0047] In this example, when obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, it is obtained at the preset frequency. The difference between the frequency and the positioning frequency of the ordinary GPS interval setting time is within the set range. In order to ensure the accuracy of positioning, the frequency is not lower than 10HZ, and in order to avoid increasing too much calculation, the positioning frequency is not higher than 20HZ.

[0048] In some optional embodiments, distance data of the environmental target in front of the collection vehicle relative to the collection vehicle is obtained, including lateral distance, longitudinal distance and angle information of the environmental target relative to the collection vehicle.

[0049] Among them, the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle is obtained, and the distance data directly obtained is the straight-line longitudinal distance data and angle data, which are converted into the lateral distance and the longitudinal distance according to the straight-line longitudinal distance data and the angle data.

[0050] In addition, when obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the environmental target in front of the collection vehicle is simplified to a data point. When the same environmental target is simplified to point data, the data at the same position of the same environmental target is used as the data point. For example, when the environmental target is a vehicle, the data at the bottom of the vehicle's centerline is taken as the data point.

[0051] Environmental targets include environmental vehicles (vehicles in front of the collection vehicle), pedestrians, and lane lines.

[0052] In some optional embodiments, obtaining distance data of an environmental target in front of a collection vehicle relative to the collection vehicle includes the following steps:

[0053] Two intelligent cameras and two millimeter-wave radars are used to collect data on the lateral collection distance and longitudinal collection distance of environmental targets relative to the collection vehicle, and an ordinary camera is used to obtain video information of environmental targets in front of the collection vehicle.

[0054] The validity of the lateral and longitudinal collection distances of the environmental targets relative to the collection vehicle collected by the two smart cameras and the two millimeter-wave radars is judged. The video information of the environmental targets in front of the collection vehicle is obtained by combining with ordinary cameras, and invalid data is filtered out and data processing is performed to obtain the lateral and longitudinal distances of the environmental targets relative to the collection vehicle.

[0055] When filtering out invalid data and processing data to obtain the lateral distance and longitudinal distance of the environmental target relative to the collection vehicle, it specifically includes:

[0056] When determining the longitudinal distance of an environmental target relative to the collection vehicle:

[0057] First, the longitudinal collection distance data of the environmental target collected by the two millimeter-wave radars relative to the collection vehicle are compared. When the difference between the longitudinal collection data collected by the two millimeter-wave radars is within a first set range, the average of the longitudinal collection data of the two millimeter-wave radars is used as the longitudinal distance of the environmental target relative to the collection vehicle.

[0058] When the difference in the longitudinal collection distance data collected by the two millimeter-wave radars is not within the first set range, the average of the longitudinal collection distance data of the environmental target collected by the two smart cameras relative to the collection vehicle is compared with the average of the longitudinal collection distance data of the environmental target collected by the two millimeter-wave radars relative to the collection vehicle, and the one with the smaller error with the average of the longitudinal collection distance data of the two smart cameras is taken as the final longitudinal distance of the environmental target relative to the collection vehicle.

[0059] When determining the lateral distance of an environmental target relative to the collection vehicle:

[0060] First, the lateral collection distance data of the environmental target of the two smart cameras relative to the collection vehicle are compared. When the difference in the lateral collection data collected by the two smart cameras is within the second set range, the average of the lateral collection data of the two smart cameras is used as the lateral distance of the environmental target relative to the collection vehicle.

[0061] When the difference in the lateral collection distance data collected by the two intelligent cameras is not within the second set range, an ordinary camera is used to obtain video information of the environmental target in front of the collection vehicle for manual re-judgment to determine the lateral distance of the environmental target relative to the collection vehicle.

[0062] After using an ordinary camera to obtain video information of the environmental target in front of the collection vehicle for manual re-judgment, the manually judged data is selected as the lateral distance of the environmental target relative to the collection vehicle.

[0063] In this example, the use of millimeter-wave radar can collect less data, avoiding the need for high-capacity storage devices. In addition, the use of millimeter-wave radar can also reduce costs.

[0064] The millimeter-wave radar is used to obtain longitudinal distance data and angle information, and the lateral distance data can be converted based on the longitudinal distance data and angle information. The two smart cameras are used to obtain the lateral distance data.

[0065] In addition, in some optional embodiments, when two millimeter-wave radars are used to collect lateral and longitudinal collection distance data about environmental targets relative to the collection vehicle, RCS (Radar Cross Section) of environmental vehicles and pedestrians in the environmental targets is also collected.

[0066] In this example, the RCS of environmental vehicles and pedestrians in the environmental targets collected by two millimeter-wave radars can be interpolated using every two adjacent RCS data to obtain the RCS of the entire trajectory environmental vehicles and pedestrians when performing database simulation reproduction.

[0067] In some optional embodiments, when two smart cameras are used to collect data on the lateral collection distance and longitudinal collection distance of the environmental target relative to the collection vehicle, contour information of environmental vehicles, pedestrians and lane lines in the environmental target is also collected and type classification is performed.

[0068] In this example, in order to facilitate data fitting, the environmental vehicles, pedestrians and lane lines are simplified into data points during data collection. However, since all of them are data points, it is impossible to distinguish the types of data points during reproduction. Therefore, when simplifying the environmental vehicles, pedestrians and lane lines into data points, the contour information of the environmental vehicles, pedestrians and lane lines in the environmental targets is collected at the same time, and the types are classified. The contour information and type classification are matched with the simplified data points of the environmental vehicles, pedestrians and lane lines, so that the environmental vehicles, pedestrians and lane lines can be reproduced when the simulation scene is established later.

[0069] like Figure 2 and Figure 3 As shown, Figure 2 The left side of the figure is a schematic diagram of converting the positioning information of the collection vehicle into global coordinates, and the right side is a schematic diagram of fitting the collection vehicle trajectory. S3: Establish a global coordinate system, fit the driving trajectory of the collection vehicle according to the GPS data, and fit the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle.

[0070] Among them, fitting the driving trajectory of the collection vehicle according to GPS data includes:

[0071] A1: Establish a global coordinate system and convert the positioning information obtained by the collection vehicle into global coordinates.

[0072] In this example, when establishing a simulation scene model, a global coordinate system needs to be established. The positioning information of the collection vehicle during driving is obtained by using ordinary GPS interval setting time and converted to the global coordinate system to prepare for the subsequent simulation scene model.

[0073] B1: Divide the global coordinates of all positioning information collected by the vehicle during driving into segments and fit the trajectory curve.

[0074] In this example, the fitting trajectory curve is calculated using a polynomial fitting calculation within the fifth order.

[0075] C1: Smooth the segmented fitting trajectory curve to obtain the driving trajectory of the collection vehicle in the global coordinate system.

[0076] In this example, the driving trajectory of the collection vehicle in the global coordinate system is used as a reference, and the coordinates of the environmental targets in the global coordinate system can be obtained by using the distances of other environmental targets relative to the collection vehicle obtained by the sensor.

[0077] Figure 3 The upper middle part shows the trajectory of the fitted environment target, and the lower part shows the schematic diagram of the fitted collection vehicle trajectory.

[0078] In addition, according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle, the trajectory of the environmental target is fitted, including:

[0079] A2: On the driving trajectory of the collection vehicle, the lateral distance and longitudinal distance of each sampling point of the environmental target relative to the driving trajectory of the collection vehicle are superimposed to obtain the global coordinate discrete point of the environmental target.

[0080] In this example, after fitting the driving trajectory of the collection vehicle in the global coordinate system, the coordinates of the collection vehicle at each sampling point of the environmental target at the same time can be obtained by interpolation or solution, and the lateral distance and longitudinal distance of each sampling point of the environmental target relative to the driving trajectory of the collection vehicle are superimposed at the same time to obtain the global coordinate discrete points of the environmental target. The sampling points are the above-mentioned data points.

[0081] In addition, the collected vehicle speed and the data collected by the millimeter-wave radar can be converted into the speed of vehicles and pedestrians in the environment target.

[0082] B2: The global coordinate discrete points of the environmental target are fitted with piecewise polynomials to obtain the piecewise trajectory of the global coordinate discrete points of the environmental target.

[0083] C2: Smooth the uneven parts of the segmented splicing of the segmented trajectory of the global coordinate discrete points of the environmental target to obtain the trajectory of the environmental target in the global coordinate system.

[0084] In some optional embodiments, when obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the relative distance data between the environmental target and the collection vehicle is also collected by using a laser radar. Based on the relative distance data between the environmental target and the collection vehicle collected by the laser radar, the absolute distance between the driving trajectory of the collection vehicle and the trajectory of the environmental target in the global coordinate system is verified.

[0085] In this example, due to the limited resolution of the low-cost laser radar, the calibration is performed by comparing the relative data of the environmental targets and the collection vehicle detected by the laser radar, and the absolute distance between the environmental targets and the collection vehicle at the same time point in the fitted global trajectory. In order to reduce the amount of calculation while ensuring the calibration accuracy, the calibration sampling point period is set to 1s-5s according to the scene complexity, 5s for simple scenes and 1s for complex scenes. When the error between the calibration point and the fitted trajectory is less than a certain error, the data is available.

[0086] S4: Determine the orientation angles of the environment vehicle and the pedestrian according to the trajectories of the environment vehicle and the pedestrian in the trajectory of the fitted environment target.

[0087] In this example, after fitting the trajectories of the collection vehicle and the environmental target in the global coordinate system, the orientation angles of the collection vehicle and the environmental target at each sampling point can be obtained by taking the derivative of .

[0088] In addition, information on the number of lanes, lane line types, and roadside buildings is obtained through ordinary cameras.

[0089] Finally, the simulation scene model can be established through the trajectory of the collection vehicle and environmental targets in the global coordinate system, the speed and orientation angle of the collection vehicle, the speed and orientation angle of environmental vehicles and pedestrians in the environmental targets, and the types (environmental vehicles, pedestrians and lane lines) and outer contours (width) of RCS and environmental targets.

[0090] In addition, the present invention also provides a scene library data generation device. The scene library data generation device includes:

[0091] A positioning mechanism, used to obtain the positioning information of the collection vehicle while it is traveling by using a common GPS interval setting time;

[0092] The environmental target collection mechanism is used to obtain the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle.

[0093] The data processing mechanism is used to establish a global coordinate system, fit the driving trajectory of the collection vehicle according to the GPS data, and fit the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle.

[0094] In this example, the positioning mechanism and environmental target acquisition mechanism include the following sensors:

[0095]

[0096] The scene library data generated by this solution system includes:

[0097]

[0098]

[0099] In addition, it also includes a storage mechanism for storing data collected by the positioning mechanism and the environmental target collection mechanism. It includes a synchronous collection platform for synchronously storing the data collected by two smart cameras, two millimeter-wave radars, a GPS and a four-line laser radar, and a video real-time storage module for storing video information collected by ordinary cameras.

[0100] In summary, this solution uses ordinary GPS to set the interval time to obtain the positioning information of the collection vehicle when it is driving; obtain the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle; establish a global coordinate system, fit the driving trajectory of the collection vehicle according to the GPS data, and fit the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle. To obtain the positioning information of the collection vehicle when it is driving, ordinary satellite positioning GPS (meter level) is used to replace the expensive inertial navigation GPS or differential GPS in the existing solution. When obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the frequency of acquisition is close to the positioning time of ordinary GPS, which can reduce the performance requirements of the collection equipment and storage equipment, thereby forming a low-cost solution. The data of the collection vehicle in the global coordinate system is obtained by fitting, and the data of the environmental target in the global coordinate system is obtained by obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, thereby obtaining the trajectory of the collection vehicle and the environmental target in the global coordinate system. In this example, an extremely low-line laser radar is used to replace the expensive high-line laser radar in the existing solution, which can also reduce the performance requirements of the collection equipment and storage equipment, thereby forming a low-cost solution.

[0101] It should be noted that, in this application, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0102] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. A method for generating scene library data, characterized in that: The following steps are involved: Use ordinary GPS interval setting time to obtain the location information of the collection vehicle when it is driving; Obtaining distance data of the environmental target in front of the collection vehicle relative to the collection vehicle; Establish a global coordinate system, fit the driving trajectory of the collection vehicle according to the GPS data, and fit the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle; The obtaining of distance data of the environmental target in front of the collection vehicle relative to the collection vehicle includes lateral distance, longitudinal distance and angle information of the environmental target relative to the collection vehicle; The step of fitting the trajectory of the environmental target according to the driving trajectory of the collection vehicle and the distance data of the environmental target relative to the collection vehicle comprises: On the driving track of the collection vehicle, the lateral distance and longitudinal distance of each sampling point of the environmental target relative to the driving track of the collection vehicle are superimposed to obtain the global coordinate discrete point of the environmental target; The global coordinate discrete points of the environmental target are fitted with the global coordinate discrete points of the environmental target through piecewise polynomials to obtain the piecewise trajectory of the global coordinate discrete points of the environmental target; The uneven parts of the segmented splicing of the segmented trajectory of the global coordinate discrete points of the environmental target are smoothed to obtain the trajectory of the environmental target in the global coordinate system.

2. The scene library data generation method according to claim 1, characterized in that: Fit the driving trajectory of the collection vehicle based on GPS data, including: Establish a global coordinate system and convert the positioning information obtained by the collection vehicle into global coordinates; The global coordinates of all positioning information collected by the vehicle during driving are divided into segments, and the trajectory curve is fitted; The segmented fitting trajectory curve is smoothed to obtain the driving trajectory of the collection vehicle in the global coordinate system.

3. The scene library data generation method according to claim 1, characterized in that: Obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle includes the following steps: Two smart cameras and two millimeter-wave radars are used to collect data on the lateral collection distance and longitudinal collection distance of environmental targets relative to the collection vehicle, and an ordinary camera is used to obtain video information of environmental targets in front of the collection vehicle; The validity of the lateral and longitudinal collection distances of the environmental targets relative to the collection vehicle collected by the two smart cameras and the two millimeter-wave radars is judged. The video information of the environmental targets in front of the collection vehicle is obtained by combining with ordinary cameras, and invalid data is filtered out and data processing is performed to obtain the lateral and longitudinal distances of the environmental targets relative to the collection vehicle.

4. The scene library data generation method according to claim 3, characterized in that: The validity of the lateral and longitudinal collection distance data of the environmental target relative to the collection vehicle collected by the two intelligent cameras and the two millimeter-wave radars is judged, and the video information of the environmental target in front of the collection vehicle is obtained by combining with the ordinary camera, and the invalid data is screened out and the data is processed to obtain the lateral and longitudinal distances of the environmental target relative to the collection vehicle, including: When determining the longitudinal distance of an environmental target relative to the collection vehicle: First, the longitudinal distance data of the environmental targets collected by the two millimeter-wave radars relative to the collection vehicle are compared. When the difference between the longitudinal data collected by the two millimeter-wave radars is within the first set range, the average of the longitudinal data collected by the two millimeter-wave radars is used as the longitudinal distance of the environmental target relative to the collection vehicle; When the difference in the longitudinal collection distance data collected by the two millimeter-wave radars is not within the first set range, the mean longitudinal collection distance data of the environmental target collected by the two smart cameras relative to the collection vehicle is compared with the mean longitudinal collection distance data of the environmental target collected by the two millimeter-wave radars relative to the collection vehicle, and the one with the smaller error with the mean longitudinal collection distance data of the two smart cameras is taken as the final longitudinal distance of the environmental target relative to the collection vehicle; When determining the lateral distance of an environmental target relative to the collection vehicle: First, compare the lateral collection distance data of the environmental targets of the two smart cameras relative to the collection vehicle. When the difference between the lateral data collected by the two smart cameras is within the second set range, the average of the lateral data collected by the two smart cameras is used as the lateral distance of the environmental target relative to the collection vehicle; When the difference in the lateral collection distance data collected by the two intelligent cameras is not within the second set range, an ordinary camera is used to obtain video information of the environmental target in front of the collection vehicle for manual re-judgment to determine the lateral distance of the environmental target relative to the collection vehicle.

5. The scene library data generation method according to claim 3, characterized in that: When two millimeter-wave radars are used to collect data on the lateral collection distance and the longitudinal collection distance of the environmental target relative to the collection vehicle, the RCS of the environmental vehicles and pedestrians in the environmental target is also collected.

6. The scene library data generation method according to claim 3, characterized in that: When two intelligent cameras are used to collect data on the lateral collection distance and longitudinal collection distance of the environmental target relative to the collection vehicle, contour information of environmental vehicles, pedestrians and lane lines in the environmental target is also collected and type classification is performed.

7. The scene library data generation method according to claim 6, characterized in that: The orientation angles of the environment vehicles and pedestrians are determined according to the trajectories of the environment vehicles and pedestrians in the trajectory of the fitted environment target.

8. The scene library data generation method according to claim 1, characterized in that: When obtaining the distance data of the environmental target in front of the collection vehicle relative to the collection vehicle, the laser radar is also used to collect the relative distance data between the environmental target and the collection vehicle. Based on the relative distance data between the environmental target and the collection vehicle collected by the laser radar, the absolute distance between the driving trajectory of the collection vehicle and the trajectory of the environmental target in the global coordinate system is verified.

Images