Vehicle test method, vehicle test system and device based on unmanned aerial vehicle, and medium
Through the UAV collecting vehicle motion information and driving behavior information, the problems of high cost and low efficiency of closed site testing systems are solved, and more efficient autonomous vehicle testing is achieved.
Patent Information
- Application Number
- CN202410983290.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-22
- Publication Date
- 2024-11-15
AI Technical Summary
The closed site testing system is cost-effective and inefficient for testing autonomous vehicles and requires long-term installation and commissioning.
UAVs are used to collect vehicle motion information and driving behavior information, and to build site maps by unmanned institutions to obtain vehicle motion information and generate driving test results, reducing dependence on closed site testing systems.
It reduces testing costs, improves testing efficiency, and reduces the installation and commissioning time of closed site testing systems.
Smart Images

Figure CN118968354A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of autonomous driving technology, and in particular to a vehicle testing method, vehicle testing system, device and readable storage medium based on a drone. Background Art
[0002] Compared with real road testing, closed-field testing has the advantages of no external interference, controllable and repeatable test conditions, and the ability to guarantee the safety of testers to the greatest extent. It is widely used in the testing and development of high-level autonomous driving functions. At present, related technologies usually use dedicated closed-field test systems to test vehicles. However, closed-field test systems are not only expensive, resulting in high testing costs, but also require a lot of time to install and debug, greatly reducing testing efficiency. Summary of the invention
[0003] The present application provides a vehicle testing method, a vehicle testing system, a device and a readable storage medium based on a drone, which solves the problems of high testing cost and low testing efficiency in the related art of using a closed field testing system to test a vehicle.
[0004] In a first aspect, the present application provides a vehicle testing system, the vehicle testing system comprising:
[0005] Test vehicles;
[0006] A drone, the drone is used to construct a map of a site image of the closed test site where the test vehicle is located to obtain a site map;
[0007] A test device is used to communicate with the drone and the test vehicle. The test device is used to obtain vehicle motion information of the test vehicle collected by the drone based on the site map and driving behavior information of the test vehicle during the driving test of the test vehicle, and generate a driving test result corresponding to the test vehicle based on the vehicle motion information and the driving behavior information.
[0008] In a second aspect, the present application further provides a vehicle testing method based on a drone, the vehicle testing method based on a drone comprising:
[0009] During the driving test of the test vehicle, the vehicle motion information of the test vehicle collected by the drone based on the site map and the driving behavior information of the test vehicle are obtained. The site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located; the driving test result corresponding to the test vehicle is determined based on the vehicle motion information and the driving behavior information.
[0010] In a third aspect, the present application further provides a testing device, the testing device comprising a memory and a processor;
[0011] The memory is used to store computer programs;
[0012] The processor is used to execute the computer program and implement the above-mentioned drone-based vehicle testing method when executing the computer program.
[0013] In a fourth aspect, the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements the drone-based vehicle testing method as described above.
[0014] The present application discloses a vehicle testing method, a vehicle testing system, a testing device and a storage medium based on a drone. By acquiring the vehicle motion information of the test vehicle collected by the drone based on the site map, and acquiring the driving behavior information of the test vehicle, and generating a driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, the vehicle motion information of the test vehicle can be collected by the drone without using a dedicated closed site testing system to collect the vehicle motion information of the test vehicle. The time spent on installing and debugging the closed site testing system can be reduced, thereby reducing the testing cost and improving the testing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0015] In order to more clearly illustrate the technical solutions of 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 some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0016] Figure 1 is a schematic diagram of a vehicle testing system provided in an embodiment of the present application;
[0017] Figure 2 It is a schematic diagram of the structure of a drone provided in an embodiment of the present application;
[0018] Figure 3 is a schematic block diagram of a drone provided in an embodiment of the present application;
[0019] Figure 4 is a site map in OpenDrive format provided in an embodiment of the present application;
[0020] Figure 5 It is a partial lane semantic map provided in an embodiment of the present application;
[0021] Figure 6 is a structural schematic diagram of a testing device provided in an embodiment of the present application;
[0022] Figure 7 is a schematic flow chart of a vehicle testing method based on a drone provided in an embodiment of the present application;
[0023] Figure 8 It is a schematic flow chart of another drone-based vehicle testing method provided in an embodiment of the present application. DETAILED DESCRIPTION
[0024] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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 creative work are within the scope of protection of this application.
[0025] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.
[0026] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0027] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.
[0028] Compared with real road testing, closed-field testing has the advantages of no external interference, controllable and repeatable test conditions, and the ability to guarantee the safety of testers to the greatest extent. It is widely used in the testing and development of high-level autonomous driving functions. At present, related technologies usually use dedicated closed-field test systems to test vehicles. However, closed-field test systems are not only expensive, resulting in high testing costs, but also require a lot of time to install and debug, greatly reducing testing efficiency.
[0029] To this end, the present application provides a vehicle testing method, vehicle testing system, device and storage medium based on drones, which can collect vehicle motion information of the test vehicle based on the site map and obtain the driving behavior information of the test vehicle, and generate the driving test results corresponding to the test vehicle according to the vehicle motion information and driving behavior information. It can realize the use of drones to collect vehicle motion information of the test vehicle, without the need to use a dedicated closed site test system to collect vehicle motion information of the test vehicle, and can reduce the time spent on installing and debugging the closed site test system, thereby reducing the test cost and improving the test efficiency. The following will explain in detail how to conduct vehicle testing based on drones.
[0030] See also Figure 1 , Figure 1 is a schematic diagram of a vehicle testing system 100 provided in an embodiment of the present application. Figure 1 As shown, the vehicle testing system 100 includes a drone 10 , a testing device 20 and a testing vehicle 30 , and the testing device 20 is communicatively connected with the drone 10 and the testing vehicle 30 .
[0031] The drone 10 is used to construct a map of the site image of the closed test site where the test vehicle 30 is located to obtain a site map.
[0032] For example, the drone 10 can hover to take aerial photos of the closed test site, and construct a map based on the site image to obtain a site map. During the driving test of the test vehicle 30, the drone 10 can hover to take aerial photos of the test vehicle 30, and perform vehicle detection and vehicle tracking on the test vehicle 30 based on the site map to obtain vehicle motion information of the test vehicle 30. Of course, the drone 10 can also fly following the test vehicle 30, perform vehicle detection and vehicle tracking on the test vehicle 30 based on the site map, and obtain vehicle motion information of the test vehicle 30.
[0033] The test device 20 can be a server or a terminal. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. The terminal can be an electronic device such as a smartphone, a tablet computer, a laptop computer, and a desktop computer.
[0034] The test device 20 may be arranged inside or outside the test vehicle 30. The test device 20 is used to obtain the vehicle motion information of the test vehicle 30 collected by the drone 10 based on the site map, and to obtain the driving behavior information of the test vehicle 30, and to generate a driving test result corresponding to the test vehicle 30 according to the vehicle motion information and the driving behavior information. It should be noted that the driving behavior information of the test vehicle 30 is collected by the test device 20.
[0035] UAV 10, specifically Figure 2 and Figure 3 As shown, it may include a body 11, a gimbal 12, a shooting device 13, a power system 14, a control system 15, and the like.
[0036] The fuselage 11 may include a fuselage and a tripod (also called a landing gear). The fuselage may include a center frame and one or more arms connected to the center frame, and the one or more arms extend radially from the center frame. The tripod is connected to the fuselage to provide support when the drone 10 lands.
[0037] The gimbal 12 is mounted on the body 11 and is used to carry a shooting device 13. The gimbal 12 may include three motors, that is, the gimbal 12 is a three-axis gimbal. Under the control of the control system 15 of the drone 10, the shooting angle of the shooting device 13 can be adjusted. The shooting angle can be understood as the angle of the direction of the lens of the shooting device 13 toward the target to be shot relative to the horizontal direction or the vertical direction.
[0038] In some embodiments, the gimbal 12 may further include a controller for controlling the movement of the gimbal 12 by controlling the motor of the gimbal, thereby adjusting the shooting angle of the shooting device 13. It should be understood that the gimbal 12 may be independent of the drone 10 or may be a part of the drone 10. It should also be understood that the motor may be a DC motor or an AC motor; or the motor may be a brushless motor or a brushed motor.
[0039] The shooting device 13 may be, for example, a device for capturing images, such as a camera or a video camera, and the shooting device 13 may communicate with the control system 15 and shoot under the control of the control system 15. In the embodiment of the present application, the shooting device 13 is mounted on the body 11 of the drone 10 via the gimbal 12. It is understandable that the shooting device 13 may also be directly fixed on the body 11 of the drone 10, so that the gimbal 12 may be omitted.
[0040] In some embodiments, the camera 13 may be controlled to shoot the closed test site at a top-down angle to obtain a site image of the closed test site. The top-down angle is when the optical axis direction of the lens of the camera 13 is perpendicular to the closed test site to be photographed, or approximately perpendicular to the closed test site to be photographed, such as 88 degrees or 92 degrees, etc., and of course other angles may also be used, which are not limited here.
[0041] In some embodiments, the shooting device 13 may include a monocular camera or a binocular camera to realize shooting of different functions. For example, a monocular camera is used to shoot the site image of the closed test site, and a binocular camera can obtain a depth image of the closed test site. The depth image includes the distance information of the target objects on the closed test site, such as other ordinary vehicles, pedestrians, and road infrastructure. The depth image can also be used as a type of traffic scene data.
[0042] The power system 14 may include one or more electronic speed regulators (referred to as ESCs), one or more propellers, and one or more motors corresponding to the one or more propellers, wherein the motor is connected between the electronic speed regulator and the propeller, and the motor and the propeller are arranged on the arm of the drone 10. The electronic speed regulator is used to receive the driving signal generated by the control system 15, and provide a driving current to the motor according to the driving signal to control the speed of the motor and then drive the propeller to rotate, thereby providing power for the flight of the drone 10, and the power enables the drone 10 to achieve one or more degrees of freedom. In some embodiments, the drone 10 can rotate around one or more rotation axes.
[0043] For example, the above-mentioned rotation axis may include a roll axis (Roll), a yaw axis (Yaw) and a pitch axis (Pitch). It should be understood that the motor may be a DC motor or an AC motor. In addition, the motor may be a brushless motor or a brushed motor.
[0044] The control system 15 may include a controller and a sensor system. The controller is used to control the flight of the drone 10. For example, the flight of the drone 10 may be controlled according to the attitude information measured by the sensor system. It should be understood that the controller may control the drone 10 according to pre-programmed instructions, or may control the drone 10 by responding to one or more control instructions from a control terminal.
[0045] The sensing system is used to measure the attitude information of the drone 10, that is, the position information and state information of the drone 10 in space, such as three-dimensional position, three-dimensional angle, three-dimensional velocity, three-dimensional acceleration and three-dimensional angular velocity.
[0046] The sensing system may include, for example, at least one of a gyroscope, an ultrasonic sensor, an electronic compass, an inertial measurement unit (IMU), a visual sensor, a global navigation satellite system, a barometer, etc. For example, the global navigation satellite system may be a global positioning system (GPS).
[0047] In the embodiment of the present application, the position information of the target object such as the test vehicle 30 in the image can be calculated by using the position information and state information of the drone 10 in space, combined with the site image taken by the drone 10. For example, according to the flight altitude of the drone 10, the field angle of the captured image and the position information of the drone, the position information of the target object in the image can be calculated through coordinate conversion, trigonometric relationship and the pixel position of the target object in the image.
[0048] In some embodiments, the drone 10 may further include a radar device, which is installed on the drone 10, specifically on the body 11 of the drone 10, and is used to measure the surrounding environment of the drone 10, such as obstacles, etc., during the flight of the drone 10 to ensure the safety of the flight.
[0049] The radar device is mounted on the tripod of the drone 10, and the radar device is connected to the control system 15 for communication. The radar device transmits the collected observation data to the control system, which is processed by the control system 15. The drone 10 may include two or more tripods, and the radar device is mounted on one of the tripods. The radar device may also be mounted at other locations of the drone 10, which is not specifically limited.
[0050] Exemplarily, the drone 10 may be a rotary-wing drone, such as a quad-rotor drone, a hexacopter drone, or an octo-rotor drone, or a fixed-wing drone, or a combination of rotary-wing and fixed-wing drones, which is not limited here.
[0051] In an embodiment of the present application, the drone 10 collects vehicle motion information of the test vehicle 30 based on the site map, and obtains driving behavior information of the test vehicle 30, and generates a driving test result corresponding to the test vehicle 30 based on the vehicle motion information and the driving behavior information. This allows the drone 10 to collect vehicle motion information of the test vehicle, without the need to use a dedicated closed field test system to collect vehicle motion information of the test vehicle 30, and can reduce the time spent on installing and debugging the closed field test system, thereby reducing testing costs and improving testing efficiency.
[0052] In some embodiments, the drone is used to map the site image based on the OpenDrive map editor or the lanlet2 map editor to obtain a site map.
[0053] In the embodiment of the present application, the construction of the map based on the site image can be completed by a drone or a test device. The following will take the construction of the map based on the site image by a drone as an example for explanation.
[0054] For example, the drone can call the OpenDrive map editor to map the site image, obtain the OpenDrive map, and determine the OpenDrive map as the site map. The specific process of map construction based on the OpenDrive map editor can be referred to in the related art, and the specific process is not limited here.
[0055] See also Figure 4 , Figure 4 It is a site map in OpenDrive format provided in an embodiment of the present application.
[0056] For example, the drone can call the lanlet2 map editor to map the site image, obtain the lanlet2 map, and determine the lanlet2 map as the site map. The specific process of map construction based on the lanlet2 map editor can be referred to in the relevant technology, and the specific process is not limited here.
[0057] In the above-mentioned embodiments, since the OpenDrive map editor and the lanlet2 map editor support multiple operations and provide a graphical interactive road network editor and a scene editor, it is convenient to construct complex road networks and traffic scenes. By constructing a map of the site image based on the OpenDrive map editor or the lanlet2 map editor, it is possible to construct complex traffic scenes, thereby improving the authenticity and applicability of the site map.
[0058] In the embodiment of the present application, in addition to constructing a site map in OpenDrive format or lanlet2 format, a lane semantic map can also be constructed. Among them, the lane semantic map may include multiple lanes and the lane line type and driving direction corresponding to each lane. The lane semantic map can subdivide the motion state of a target such as a test vehicle by lane line, lane line type, and lane line driving direction, which can effectively improve the accuracy of the vehicle motion information of the subsequent detection test vehicle.
[0059] In some embodiments, the drone is used to: perform lane line detection on the site image to obtain lane line detection results of the closed test site, the lane line detection results include a multi-frame lane line semantic map and a multi-frame centerline semantic map; perform lane positioning based on the multi-frame lane line semantic map and the multi-frame centerline semantic map to obtain a lane semantic map of the closed test site; determine the site map based on the lane semantic map.
[0060] Specifically, based on the trained lane line detection model, lane line detection can be performed on the site image to obtain the lane line detection results of the closed test site, and the lane line detection results include multi-frame lane line semantic maps and multi-frame center line semantic maps. The lane line semantic map may include information such as lane line type, lane line angle, lane line offset, etc., and the center line semantic map may include information such as center line type, center line angle, and center line offset.
[0061] The lane line detection model may be a pre-trained model, and the lane line detection model is used to detect lane lines and center lines, and output a lane line semantic map and a center line semantic map.
[0062] In an embodiment of the present application, the lane detection model may include a U-Net3+ network model, and the U-Net3+ network model includes multiple encoding layers and multiple decoding layers. It should be noted that the U-Net3+ network model adopts a symmetrical U-shaped structure that fused downsampling and upsampling, and uses semantic segmentation to process the input image to finally form a semantic segmentation map. Among them, the decoding layer of the penultimate layer is configured as an output layer; the output layer may include a lane probability output layer, a lane type output layer, a lane angle output layer, a lane offset output layer, a lane centerline probability output layer, a centerline type output layer, a centerline angle output layer, and a centerline offset output layer.
[0063] Specifically, performing lane positioning according to a multi-frame lane line semantic map and a multi-frame centerline semantic map to obtain a lane semantic map of a closed test site may include: performing lane line recognition on the multi-frame lane line semantic map to obtain a lane line recognition map; performing centerline recognition on the multi-frame centerline semantic map to obtain a centerline recognition map; performing lane positioning according to the lane line recognition map and the centerline recognition map to obtain the lane position of each lane in the closed test site; and drawing lane areas according to the lane position of each lane and the lane line type to obtain a lane semantic map.
[0064] Among them, performing lane line recognition on multiple frames of lane line semantic maps to obtain a lane line recognition map can include: stacking multiple frames of lane line semantic maps to obtain a lane line semantic segmentation map; binarizing the lane line semantic segmentation map to obtain a binary lane line semantic segmentation map; performing connected domain analysis on the binary lane line semantic segmentation map to obtain a connected domain analysis result map, which includes multiple connected domains; performing lane line recognition on the connected domain analysis result map to obtain a lane line recognition map.
[0065] It should be noted that the process of center line recognition is similar to that of lane line recognition and will not be described in detail here.
[0066] Exemplarily, the lane line semantic segmentation map can be binarized to obtain a binarized lane line semantic segmentation map. The binarization process is used to binarize the stacked lane line semantic segmentation map to strengthen the representation of the lane line. The specific operation is to set the position with a confidence greater than 0 in the lane line semantic segmentation map to 1, and set the position with a confidence less than 0 to 0, so that the lane line and the background in the lane line semantic segmentation map can be clearly distinguished.
[0067] For example, a connected domain analysis can be performed on the binary lane semantic segmentation map to obtain a connected domain analysis result map, which includes multiple connected domains. It should be noted that a connected domain refers to a set of all connected pixels with the same pixel value in a binary image. The steps of the connected domain analysis are as follows:
[0068] (1) Initialize the labeling matrix: Create a matrix with the same size as the lane semantic segmentation map, with all elements set to 0, to record the connected domain number of each pixel. Initially, set the connected domain number i to 1.
[0069] (2) Pixel-by-pixel scanning and marking:
[0070] a. If a pixel is not marked around it (i.e., all around it are 0), it is marked as a new connected domain, numbered i, and i is incremented by 1.
[0071] b. If all labeled points in the neighborhood have the same label, the point inherits this label.
[0072] c. If there are different labels in the neighborhood, select the smallest label and record the equivalence relationship of all different labels.
[0073] (3) Second scan: Rescan each pixel point and update the labels of all equivalent connected domains to ensure that all equivalent connected domains use the smallest connected domain number. This can improve the uniformity and accuracy of connected domain labeling.
[0074] (4) Generate connected domains: Based on the final connected domain labels, each independent connected domain can be extracted from the lane line semantic segmentation map.
[0075] In the above embodiment, by performing connected domain analysis on the binary lane line semantic segmentation map, independent objects or regions in the lane line semantic segmentation map can be identified, which helps to accurately identify individual lane lines in complex road images.
[0076] Specifically, performing lane line recognition on a connected domain analysis result graph to obtain a lane line recognition graph may include: performing adjacent point matching on each connected domain in the filtered connected domain analysis result graph to obtain a connected domain analysis result graph after adjacent point matching; performing connected domain connection on the connected domain analysis result graph after adjacent point matching to obtain a connected domain analysis result graph after connected domains; performing polynomial fitting on the connected domains in the connected domain analysis result graph after connected domains to obtain a lane line recognition graph.
[0077] Exemplarily, when matching adjacent points for each connected domain in the filtered connected domain analysis result graph, the following steps can be performed for each connected domain: First, scan the connected domain from left to right, find the first point whose pixel value is not zero, and calculate its angle. Subsequently, mark the point as processed and set its pixel value to zero. Next, search for adjacent non-zero pixels in the surrounding area centered on the point (expanded by 1 pixel). For each adjacent point found, calculate the angle difference between it and the current point, and select the point with the smallest angle difference as the matching point. At the same time, mark the matching point as processed and set its pixel value to zero. Repeat this process until all connected domains are searched.
[0078] Among them, the connected domain analysis result graph after the adjacent points are matched is connected to obtain the connected domain analysis result graph after the connected domains are connected, including: determining the distance between two adjacent connected domains in the connected domain analysis result graph; if there are two adjacent connected domains whose distance is less than a preset distance value, the two adjacent connected domains are connected.
[0079] Exemplarily, the distance between two adjacent connected domains may be calculated, and for two adjacent connected domains whose distance is less than a preset distance value, the two adjacent connected domains are connected. The preset distance value may be set according to actual conditions, and the specific value is not limited here.
[0080] In the above embodiment, by connecting two adjacent connected domains whose distance is less than a preset distance value, the broken parts of the lane line can be filled, which can not only enhance the integrity of the lane line, but also better handle the lane line in complex road scenes, thereby improving the accuracy and stability of lane line detection.
[0081] Exemplarily, a polynomial fitting may be performed on the connected domains in the connected domain analysis result graph after the connected domains are connected to obtain a lane line recognition graph. For example, a cubic polynomial fitting may be performed on the connected domains in the connected domain analysis result graph after the connected domains are connected, wherein the specific fitting process may refer to the detailed description of the related technology and will not be described in detail here.
[0082] Exemplarily, the lane line recognition map includes a lane line offset corresponding to the lane line, and the center line recognition map includes a center line offset corresponding to the center line. The lane line offset refers to the shortest distance between the lane line and the center of the image in the image, which can be understood as the vertical distance between the center of the image and the lane line. In the embodiment of the present application, the positional relationship between the lane line and the center line in the image can be determined based on the lane line offset, and then the upper and lower lane lines of a center line can be determined.
[0083] In some embodiments, lane positioning is performed based on a lane line recognition map and a center line recognition map to obtain lane positions of each lane in a closed test site, which may include: extracting at least one lane line in the lane line recognition map, and extracting at least one center line in the center line recognition map; performing lane line matching on each center line based on a lane line offset corresponding to each lane line and a center line offset corresponding to each center line, and determining the lane position of each lane based on the lane line matched to each center line.
[0084] For example, since the lane lines must be located on both sides of the center line, each center line can be processed in turn, and the upper boundary of the lane can be searched upward along the points on the center line according to the center line offset of the center line, and the lower boundary of the lane can be searched downward, so that the lane position of the lane can be quickly and accurately located.
[0085] After determining the lane positions of each lane in the closed test site, the lane area can be drawn according to the lane positions and lane line types of each lane to obtain a lane semantic map. Figure 5 As shown, Figure 5 It is a partial lane semantic map provided by the embodiment of the present application, which can divide and number each lane in the lane semantic map. Among them, the lane semantic map may include information such as lane lines, lane line types, lane line driving directions, center lines, center line types, etc.
[0086] Exemplarily, after the lane semantic map is obtained, the lane semantic map may be determined as the site map.
[0087] In the above embodiment, lane line detection is performed on the site image, and lane positioning is performed based on the multi-frame lane line semantic map and the multi-frame center line semantic map obtained by detection, so as to obtain a lane semantic map of the closed test site. Since the lane semantic map can subdivide the running status of targets such as vehicles by lane lines, lane line types, and driving directions of lane lines, it can effectively improve the accuracy of vehicle motion information of subsequent detection test vehicles.
[0088] In some embodiments, the drone is also used to: perform vehicle detection and vehicle tracking on the test vehicle based on the site map to obtain vehicle trajectory data of the test vehicle; determine vehicle characteristic parameters of the test vehicle based on the vehicle trajectory data; and generate vehicle motion information based on the vehicle trajectory data and the vehicle characteristic parameters.
[0089] Specifically, performing vehicle detection and vehicle tracking on the test vehicle based on the site map to obtain vehicle trajectory data of the test vehicle may include: performing vehicle detection on the site map to obtain vehicle static information of the test vehicle; and performing vehicle tracking according to the vehicle static information of the test vehicle to obtain vehicle trajectory data of the test vehicle.
[0090] In an embodiment of the present application, a trained target detection model can be used to detect vehicles on a site map. The target detection model may include, but is not limited to, a Mask R-CNN (Faster R-CNN with added mask layer) model, a YOLO (You Only Look Once) model, an SSD (Single Shot MultiBox Detector) model, and the like. For example, the Mask R-CNN model can be used to detect vehicles on a site map to obtain static vehicle information of a test vehicle.
[0091] Exemplarily, the static information of a vehicle may include location information, vehicle profile information, vehicle identification information, vehicle type, etc. Among them, the location information may include the location coordinates of the vehicle on the road. The vehicle profile information may include the vehicle's external dimensions, such as the vehicle length and width. The vehicle identification information refers to the vehicle's ID (IDentity, identification code), for example, the vehicle may be numbered to obtain the vehicle's ID. The vehicle type may include several major categories such as cars, buses, and trucks.
[0092] It should be noted that since the vehicle static information reflects the vehicle position and status of the test vehicle at a certain moment, but cannot reflect the dynamic changes of the test vehicle over time, it is necessary to further track the test vehicle to obtain the vehicle dynamic information of the test vehicle, that is, the vehicle driving trajectory data.
[0093] In some embodiments, vehicle tracking is performed based on the static vehicle information corresponding to the test vehicle to obtain vehicle dynamic information corresponding to the test vehicle, including: inputting the test vehicle's location information, vehicle profile information, vehicle identification information, and vehicle type into a trained multi-target tracking model to perform vehicle tracking and obtain vehicle driving trajectory data corresponding to the test vehicle.
[0094] Among them, the multi-target tracking model can be trained using the SORT (Simple Online and Realtime Tracking) algorithm to track vehicles. It should be noted that the SORT algorithm is a multi-target tracking algorithm based on the Kalman filter and the Hungarian algorithm, which can handle the multi-target tracking problem in the video in real time. The main idea of the SORT algorithm is to use the Kalman filter algorithm to predict the motion state of the target, and use the IoU (Intersection over Union) algorithm to calculate the similarity between the detection box and the tracking box, and finally use the Hungarian algorithm to find the best matching result.
[0095] In some embodiments, the vehicle driving trajectory data may include the lane, position, speed, and acceleration corresponding to the test vehicle at different times.
[0096] Exemplarily, after obtaining the vehicle trajectory data of the test vehicle, the vehicle characteristic parameters of the test vehicle can be determined according to the vehicle trajectory data. The vehicle characteristic parameters include at least one of the following: collision time, headway, headway, vehicle yaw rate, vehicle offset, collision deceleration rate. For example, the vehicle characteristic parameters may include collision time and headway.
[0097] The time-to-collision (TTC) indicates the time it takes for the main vehicle to collide with the vehicle in front without taking any collision avoidance measures. The calculation formula is as follows:
[0098]
[0099] In the formula, x(t) and v(t) are the position and velocity of the vehicle at time t, respectively. The subscripts p and ego represent the preceding vehicle and the ego vehicle, respectively. p is the length of the preceding vehicle. The larger the TTC, the longer the collision time, that is, the smaller the probability of collision between the two vehicles. On the contrary, the smaller the collision time, the greater the probability of collision between the two vehicles.
[0100] Time Headway (THW) refers to the time required for the main vehicle to reach the current position of the preceding vehicle at the current speed. The calculation formula is as follows:
[0101]
[0102] Where D HW is the distance between the vehicle and the front vehicle, is the speed of the vehicle, D HW The larger the value is, the greater the following distance between the two vehicles is, and the probability of collision between the two vehicles is small. Conversely, the probability of collision between the two vehicles is large.
[0103] The calculation formula for distance headway (DHW) is as follows:
[0104] D HW =x p (t)-x ego (t) (3)
[0105] In the formula, x(t) represents the position of the vehicle at time t. The headway distance refers to the distance from the front of the vehicle to the front of the vehicle when the vehicle is traveling at a constant speed in the current state. The larger the DHW, the greater the distance between the vehicle and the front vehicle, and the smaller the probability of a collision between the two vehicles, and vice versa.
[0106] The vehicle yaw rate refers to the rotation speed of the vehicle around the vertical axis, which can be calculated by the vehicle driving direction and lane direction in the lane semantic map. The calculation formula is as follows:
[0107]
[0108] Where Δθ represents the angular difference between the vehicle's driving direction and the lane direction, and t represents the time when the angle changes. The vehicle's yaw rate can be used to more accurately assess the degree to which the vehicle deviates from the center line, which helps to establish the vehicle motion model.
[0109] The ego offset refers to the offset between the center point of the vehicle and the center line of the lane, which is a maximum of half the lane width and can be directly obtained from the lane semantic map.
[0110] The Deceleration Rate to Avoid a Crash (DRAC) indicates the deceleration required by the vehicle to avoid a crash in the current state. The calculation formula is as follows:
[0111]
[0112] Among them, v represents the speed of the vehicle, x represents the position of the vehicle, and l represents the length of the vehicle. The larger the DRAC is, the greater the required collision avoidance deceleration is, the greater the relative speed of the two vehicles is, and the higher the probability of collision between the two vehicles is. Otherwise, the probability of collision between the two vehicles is low.
[0113] In the embodiment of the present application, the vehicle characteristic parameters can be calculated according to the vehicle trajectory data based on the above calculation formula. The specific calculation process is not described in detail here.
[0114] For example, after calculating the vehicle characteristic parameters of the test vehicle according to the vehicle trajectory data, the vehicle trajectory data and the vehicle characteristic parameters can be determined as vehicle motion information. The vehicle motion information can include the lane, position, speed, acceleration corresponding to the test vehicle at different times, and the collision time, headway, headway, vehicle yaw rate, self-vehicle offset, collision deceleration rate and other parameters corresponding to the test vehicle.
[0115] The above embodiment obtains vehicle trajectory data of the test vehicle by performing vehicle detection and vehicle tracking on the test vehicle based on the site map, determines vehicle characteristic parameters of the test vehicle according to the vehicle trajectory data, and generates vehicle motion information according to the vehicle trajectory data and the vehicle characteristic parameters. This makes it possible to collect vehicle motion information of the test vehicle by using a drone, without using a dedicated closed site test system to collect vehicle motion information of the test vehicle. This can reduce the time spent on installing and debugging the closed site test system, thereby reducing testing costs and improving testing efficiency.
[0116] In some embodiments, the test device is used to: collect driver behavior information of a driver driving a test vehicle, and collect vehicle status information of the test vehicle; and generate driving behavior information based on the driver behavior information and the vehicle status information.
[0117] In an embodiment of the present application, a photographing device is configured in the test vehicle, and the photographing device is communicatively connected with the test device. The photographing device is used to photograph a first image including the driver; the test device is used to obtain the first image and determine the driver behavior information according to the first image.
[0118] The driver behavior information includes the driver's operating behavior on the test vehicle and the operating time corresponding to the operating behavior. For example, the operating behavior may include turning the steering wheel, stepping on the accelerator pedal, stepping on the brake pedal, pulling the handbrake, turning on the turn signal, etc.
[0119] In the embodiment of the present application, human posture recognition technology can be used to detect the driver's operating behavior based on the first image. The specific detection process is not limited here. Of course, the driver's operating behavior and the operating time corresponding to the operating behavior can also be determined by detecting the electrical signals generated by the steering wheel, accelerator pedal, brake pedal, handbrake, and turn signal.
[0120] Exemplarily, the test device may also collect vehicle status information of the test vehicle.
[0121] In the embodiment of the present application, a bus data acquisition tool is configured in the test vehicle, and the bus data acquisition tool is connected to the test device in communication, and the test device is used to collect vehicle status information of the test vehicle through the bus data acquisition tool. The bus data acquisition tool can be a DEWETRON series data acquisition tool or other types of data acquisition tools, which are not limited here.
[0122] Exemplarily, the vehicle status information includes at least one of the following: steering wheel rotation angle, accelerator pedal amplitude, brake pedal amplitude. For example, the vehicle status information may include steering wheel rotation angle, accelerator pedal amplitude, and brake pedal amplitude, and may also include other information, which is not limited here.
[0123] In the above-mentioned embodiment, the driver behavior information of the driver driving the test vehicle is collected based on the photographing device, and the vehicle status information of the test vehicle is collected through the bus data collection tool. The operation and detection are convenient, and the driving behavior information of the test vehicle can be obtained quickly. There is no need to use a dedicated closed field test system to collect the driving behavior information of the test vehicle, which can reduce the test cost and improve the test efficiency.
[0124] See also Figure 6 , Figure 6 2 is a schematic diagram of a test device 20 provided in an embodiment of the present application. The test device 20 may include a processor 201 and a memory 202, wherein the processor 201 and the memory 202 may be connected via a bus, and the bus may be any applicable bus such as an Inter-integrated Circuit (I2C) bus.
[0125] Exemplarily, the test device 20 can be a server or a terminal. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms. The terminal can be an electronic device such as a smart phone, a tablet computer, a laptop computer, and a desktop computer.
[0126] Exemplarily, the memory 202 may include a storage medium and an internal memory. The storage medium may store an operating system and a computer program. The computer program includes program instructions, which, when executed, may enable the processor 201 to execute the vehicle testing method based on a drone described in any embodiment.
[0127] The processor 201 is used to provide computing and control capabilities to support the operation of the entire testing device 20 .
[0128] The processor 201 may be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other types of processors such as programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components. A general-purpose processor may be a microprocessor, or a general-purpose processor may be any conventional processor.
[0129] In some implementations, the processor 201 is used to run a computer program stored in the memory 202 to implement the following steps:
[0130] During the driving test of the test vehicle, the drone collects vehicle motion information of the test vehicle based on the site map, and obtains the driving behavior information of the test vehicle. The site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located; the driving test result corresponding to the test vehicle is determined based on the vehicle motion information and the driving behavior information.
[0131] In one embodiment, when the processor 201 determines the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, it is used to implement:
[0132] The vehicle motion information is forward filtered and backward filtered to obtain filtered vehicle motion information; and a driving test result is generated according to the driving behavior information and the filtered vehicle motion information.
[0133] In one embodiment, when the processor 201 determines the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, it is used to implement:
[0134] When the coordinate system used by the vehicle motion information is different from the coordinate system used by the site image, the vehicle motion information is aligned according to the coordinate system used by the site image to obtain the vehicle motion information after the coordinate system is aligned; and the driving test result is generated based on the driving behavior information and the vehicle motion information after the coordinate system is aligned.
[0135] In one embodiment, after determining the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, the processor 201 is used to implement:
[0136] Acquire multiple frames of site images of a closed test site collected by a drone; generate road traffic images based on the multiple frames of site images; and visualize the driving test results in a preset visualization interface based on the road traffic images.
[0137] See also Figure 7 , Figure 7 FIG. 1 is a schematic flow chart of a vehicle testing method based on a drone provided in an embodiment of the present application. Figure 7 As shown, the drone-based vehicle testing method may include step S301 and step S302.
[0138] Step S301: During the driving test of the test vehicle, the drone collects vehicle motion information of the test vehicle based on a site map, and obtains driving behavior information of the test vehicle. The site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located.
[0139] For example, during the driving test of the test vehicle, the test device can obtain the vehicle motion information of the test vehicle collected by the drone based on the site map, wherein the site map is obtained by mapping the site image of the closed test site where the test vehicle is located by the drone.
[0140] In the embodiment of the present application, the site map can be obtained by mapping the site image by a drone, or by mapping the site image by a test device. The specific process of the test device constructing the site map is the same as the specific process of the drone constructing the site map, which can be referred to the detailed description of the above embodiment and will not be repeated here.
[0141] Exemplarily, during the driving test of the test vehicle, the test device may also obtain driving behavior information of the test vehicle.
[0142] Specifically, the test device can collect driver behavior information of the driver driving the test vehicle, and collect vehicle status information of the test vehicle, and generate driving behavior information based on the driver behavior information and vehicle status information. The specific process of collecting driver behavior information and vehicle status information can be referred to the detailed description of the above embodiment, and the specific process is not repeated here.
[0143] Step S302: Determine the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information.
[0144] Exemplarily, after obtaining the vehicle motion information and driving behavior information of the test vehicle, a driving test result corresponding to the test vehicle can be generated according to the vehicle motion information and driving behavior information. The driving test result can include vehicle driving trajectory data, vehicle characteristic parameters, driver behavior information, and vehicle status information.
[0145] It should be noted that by generating driving test results corresponding to the test vehicle based on vehicle motion information and driving behavior information, more accurate and reliable structured data can be provided for the development and test verification of the autonomous driving system.
[0146] In the above embodiment, by acquiring vehicle motion information of the test vehicle based on the site map collected by the drone, and acquiring the driving behavior information of the test vehicle, and generating a driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, it is possible to utilize the drone to collect the vehicle motion information of the test vehicle, without the need to use a dedicated closed field test system to collect the vehicle motion information of the test vehicle, and the time spent on installing and debugging the closed field test system can be reduced, thereby reducing the test cost and improving the test efficiency.
[0147] In some embodiments, determining a driving test result corresponding to a test vehicle based on vehicle motion information and driving behavior information may include: forward filtering and backward filtering the vehicle motion information to obtain filtered vehicle motion information; and generating a driving test result based on the driving behavior information and the filtered vehicle motion information.
[0148] It should be noted that the vehicle motion information includes vehicle driving trajectory data and vehicle characteristic parameters, among which the position stability of the trajectory points in the vehicle driving trajectory data has a greater impact on the data accuracy. The trajectory points are generally the center position of the detection frame. In order to further improve the quality of the vehicle driving trajectory data and the accuracy of the vehicle motion information, the embodiment of the present application can smooth the vehicle motion information.
[0149] Exemplarily, the RTS (Rauch-Tung-Striebel smoother) algorithm can be used to perform forward filtering and backward filtering on the vehicle motion information of the test vehicle to obtain the filtered vehicle motion information of the test vehicle. The specific process of forward filtering and backward filtering can be referred to the relevant technology and is not limited here. Of course, other types of smoothing algorithms can also be used to smooth the vehicle motion information, and the embodiment of the present application does not limit the type of smoothing algorithm.
[0150] It should be noted that the RTS algorithm is a fixed interval smoother based on the maximum likelihood estimation criterion. The RTS smoothing algorithm is based on the Kalman filter and includes two stages: forward filtering and backward filtering. Among them, the forward filtering adopts the traditional Kalman filter, while the backward filtering reuses the data on this basis to improve the accuracy of state estimation.
[0151] In the above embodiment, by using the RTS algorithm to perform forward filtering and backward filtering on the vehicle motion information of the test vehicle, the accuracy of the speed and acceleration in the vehicle driving trajectory data can be improved, thereby significantly improving the accuracy and reliability of the vehicle motion information.
[0152] In some embodiments, a driving test result corresponding to a test vehicle is determined based on the vehicle motion information and the driving behavior information, including: when a coordinate system used by the vehicle motion information is different from a coordinate system used by the site image, aligning the vehicle motion information according to the coordinate system used by the site image to obtain the vehicle motion information after the coordinate system is aligned; generating a driving test result based on the driving behavior information and the vehicle motion information after the coordinate system is aligned.
[0153] In the embodiment of the present application, since the site map is one of the lane semantic map, OpenDrive map and lanlet2 map, and the coordinate system used by the lane semantic map is the same as the coordinate system used by the site image, and the coordinate systems used by the OpenDrive map and the lanlet2 map are different from the coordinate system used by the site image, if vehicle detection and vehicle tracking are performed on the test vehicle based on the OpenDrive map or the lanlet2 map, the coordinate system used by the obtained vehicle motion information is different from the coordinate system used by the site image. At this time, the vehicle motion information needs to be aligned with the coordinate system.
[0154] Specifically, when the coordinate system of the vehicle motion information is aligned according to the coordinate system adopted by the site image, the coordinate system of the vehicle motion information can be aligned based on the transformation matrix so that the coordinate system adopted by the vehicle motion information after the coordinate system alignment is the same as the coordinate system adopted by the site image. The transformation matrix is as follows:
[0155]
[0156] In the formula, (x′, y′) represents the coordinates after the coordinate system is aligned; (x, y) represents the coordinates before the coordinate system is aligned. The parameters a, b, c, e, d, and f in the transformation matrix can be solved using the three pairs of coordinates in the coordinate system used by the vehicle motion information and the corresponding three pairs of coordinates in the coordinate system used by the site image.
[0157] In the above embodiment, when the coordinate system used by the vehicle motion information is different from the coordinate system used by the site image, the coordinate system used by the vehicle motion information is aligned according to the coordinate system used by the site image. This can make the coordinate system used by the vehicle motion information the same as the coordinate system used by the site image, thereby ensuring that the vehicle motion information can be accurately visualized in the road traffic image later, avoiding large position deviations when visualizing the vehicle motion information in the road traffic image.
[0158] In other embodiments, the driving test result corresponding to the test vehicle is determined based on the vehicle motion information and the driving behavior information, including: when the coordinate system used by the filtered vehicle motion information is different from the coordinate system used by the site image, the filtered vehicle motion information is aligned according to the coordinate system used by the site image to obtain the vehicle motion information after the coordinate system is aligned; and the driving test result is generated based on the driving behavior information and the vehicle motion information after the coordinate system is aligned. The specific process of aligning the filtered vehicle motion information can be referred to the detailed description of the above embodiment, which will not be repeated here.
[0159] It should be noted that by aligning the coordinate system of the filtered vehicle motion information, not only can the accuracy and reliability of the vehicle motion information be significantly improved, but also large position deviations can be avoided when marking the vehicle motion information to the road traffic image, thereby improving the accuracy of the subsequent visualization of the vehicle motion information to the road traffic image.
[0160] See also Figure 8 , Figure 8 FIG. 1 is a schematic flow chart of another vehicle testing method based on a drone provided in an embodiment of the present application. Figure 8 As shown, the drone-based vehicle testing method may include steps S401 to S405.
[0161] Step S401: During the driving test of the test vehicle, the drone collects vehicle motion information of the test vehicle based on a site map, and obtains driving behavior information of the test vehicle. The site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located.
[0162] Step S402: Determine the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information.
[0163] It can be understood that steps S401 to S402 are the same as the above-mentioned steps S301 to S302, and are not described in detail here.
[0164] Step S403: obtaining multiple frames of site images of the closed test site captured by the drone.
[0165] Exemplarily, the testing device may receive multiple frames of site images of a closed testing site captured by a drone.
[0166] Step S404: Generate a road traffic image based on the multiple frames of site images.
[0167] For example, when a road traffic image is generated based on multiple frames of site images, the multiple frames of site images may be superimposed and then an average of pixel values may be calculated to obtain the road traffic image.
[0168] It should be noted that when the UAV is hovering or moving to take aerial photos of the site, both the UAV and the test vehicle have a certain speed. Therefore, the single-frame site image taken will have dynamic noise. In order to eliminate the dynamic noise, it is necessary to superimpose multiple frames of the site image and calculate the average pixel value to eliminate the dynamic noise.
[0169] Step S405: Visualize the driving test results on a preset visualization interface based on the road traffic image.
[0170] In an embodiment of the present application, after obtaining the driving test results of the test vehicle, the driving test results can also be visualized, so that the driving test results can be intuitively displayed, which is convenient for observing and analyzing the driving test results.
[0171] Exemplarily, the vehicle motion information includes vehicle driving trajectory data and vehicle characteristic parameters, and the driving behavior information includes driver behavior information and vehicle status information; the visualization interface may include an in-vehicle video display window, a data chart window, and a vehicle trajectory window. In some embodiments, based on the road traffic image, the driving test results are visualized in a preset visualization interface, which may include at least one of the following: visualizing the driver behavior information in the in-vehicle video display window, visualizing the vehicle characteristic parameters and / or vehicle status information in a preset data chart window, and visualizing the road traffic image and vehicle driving trajectory data in a preset vehicle trajectory window.
[0172] In an embodiment of the present application, a visualization tool or a visualization system can be used to realize visualization of driving test results. It should be noted that the visualization tool or the visualization system supports the reproduction of the vehicle driving trajectory data of the test vehicle during the vehicle test in the road traffic image, supports the observation of the driving behavior of the driver during the test, and supports the display of the collected data information of interest (such as vehicle characteristic parameters, vehicle status information, etc.) in the form of charts.
[0173] For example, the vehicle trajectory data can be loaded in the vehicle trajectory window with the road traffic image as the base map, for example, the shape, size, position, speed and other parameters of the test vehicle at each moment. At a certain moment, in the road traffic image, a simple vehicle rectangle is drawn according to the position and shape of the test vehicle to represent the visualization of the test vehicle at this moment, and the vehicle trajectories of the test vehicle at all moments are played in a loop to complete the visualization of the vehicle trajectories.
[0174] For example, a video containing the driver's operating behavior of the test vehicle can be played in a video display window in the vehicle. Vehicle characteristic parameters (e.g., collision time, headway, headway, etc.) and / or vehicle status information (e.g., steering wheel rotation angle, accelerator pedal amplitude, brake plate amplitude, etc.) of the test vehicle can be displayed in a chart form in a data chart window.
[0175] The above-mentioned embodiment can realize visualization of the driving test results by displaying the driving test results on a visualization interface, and can intuitively display the vehicle motion information and driver behavior of the test vehicle during the driving test process, which is convenient for observation and analysis.
[0176] A computer-readable storage medium is also provided in an embodiment of the present application. The computer-readable storage medium stores a computer program. The computer program includes program instructions. The processor executes the program instructions to implement any one of the drone-based vehicle testing methods provided in the embodiments of the present application.
[0177] For example, the computer program is loaded by a processor and may execute the following steps:
[0178] During the driving test of the test vehicle, the drone collects vehicle motion information of the test vehicle based on the site map, and obtains the driving behavior information of the test vehicle. The site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located; the driving test result corresponding to the test vehicle is determined based on the vehicle motion information and the driving behavior information.
[0179] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.
[0180] The computer-readable storage medium may be an internal storage unit of the test device of the aforementioned embodiment, such as a hard disk or memory of the test device. The computer-readable storage medium may also be an external storage device of the test device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital card (SecureDigital, SD), a flash card (Flash Card), etc. equipped on the test device.
[0181] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.
Claims
1. A vehicle testing system, characterized in that: The vehicle testing system comprises: Test vehicles; A drone, the drone is used to construct a map of a site image of the closed test site where the test vehicle is located to obtain a site map; A test device is used to communicate with the drone and the test vehicle. The test device is used to obtain vehicle motion information of the test vehicle collected by the drone based on the site map and driving behavior information of the test vehicle during the driving test of the test vehicle, and generate a driving test result corresponding to the test vehicle based on the vehicle motion information and the driving behavior information.
2. The vehicle testing system according to claim 1, characterized in that: The drone is used to construct a map of the site image based on an OpenDrive map editor or a lanlet2 map editor to obtain the site map.
3. The vehicle testing system according to claim 1, characterized in that: The drone is used for: Performing lane line detection on the site image to obtain a lane line detection result of the closed test site, wherein the lane line detection result includes a multi-frame lane line semantic map and a multi-frame center line semantic map; Performing lane positioning according to the multiple frames of lane line semantic maps and the multiple frames of center line semantic maps to obtain a lane semantic map of the closed test site; The site map is determined according to the lane semantic map.
4. The vehicle testing system according to claim 1, characterized in that: The drone is used for: Performing vehicle detection and vehicle tracking on the test vehicle based on the site map to obtain vehicle trajectory data of the test vehicle; determining vehicle characteristic parameters of the test vehicle according to the vehicle trajectory data; The vehicle motion information is generated according to the vehicle trajectory data and the vehicle characteristic parameters.
5. The vehicle testing system according to claim 4, characterized in that: The vehicle driving trajectory data includes the lane, position, speed, and acceleration corresponding to the test vehicle at different times.
6. The vehicle testing system according to claim 4, characterized in that: The vehicle characteristic parameters include at least one of the following: collision time, headway, headway, vehicle yaw rate, vehicle offset, and collision deceleration rate.
7. The vehicle testing system according to claim 1, characterized in that: The test device is used for: Collecting driver behavior information of a driver driving the test vehicle, and collecting vehicle status information of the test vehicle; The driving behavior information is generated according to the driver behavior information and the vehicle state information.
8. The vehicle testing system according to claim 7, characterized in that: A photographing device is disposed in the test vehicle, and the photographing device is communicatively connected with the test device; The photographing device is used to photograph a first image including the driver; The testing device is used to acquire the first image and determine the driver behavior information based on the first image, wherein the driver behavior information includes the driver's operating behavior on the test vehicle and the operating time corresponding to the operating behavior.
9. The vehicle testing system according to claim 7, characterized in that: A bus data acquisition tool is configured in the test vehicle, and the bus data acquisition tool is communicatively connected with the test device. The test device is used to acquire vehicle status information of the test vehicle through the bus data acquisition tool.
10. A vehicle testing method based on drone, characterized in that: The method comprises: During the driving test of the test vehicle, the vehicle motion information of the test vehicle collected by the drone based on the site map and the driving behavior information of the test vehicle are obtained, wherein the site map is obtained by the drone mapping the site image of the closed test site where the test vehicle is located; A driving test result corresponding to the test vehicle is determined according to the vehicle motion information and the driving behavior information.
11. The vehicle testing method according to claim 10, characterized in that: The determining, according to the vehicle motion information and the driving behavior information, a driving test result corresponding to the test vehicle comprises: Performing forward filtering and backward filtering on the vehicle motion information to obtain filtered vehicle motion information; The driving test result is generated according to the driving behavior information and the filtered vehicle motion information.
12. The vehicle testing method according to claim 10, characterized in that: The determining, according to the vehicle motion information and the driving behavior information, a driving test result corresponding to the test vehicle comprises: When the coordinate system used by the vehicle motion information is different from the coordinate system used by the site image, aligning the vehicle motion information according to the coordinate system used by the site image to obtain the vehicle motion information after the coordinate system is aligned; The driving test result is generated according to the vehicle motion information after the driving behavior information is aligned with the coordinate system.
13. The vehicle testing method according to claim 10, characterized in that: After determining the driving test result corresponding to the test vehicle according to the vehicle motion information and the driving behavior information, the method further includes: Acquire multiple frames of site images of the closed test site captured by the drone; generating a road traffic image based on a plurality of frames of the site image; Based on the road traffic image, the driving test result is visualized on a preset visualization interface.
14. A testing device, characterized in that: The testing device comprises a processor and a memory; The memory is used to store computer programs; The processor is used to execute the computer program and implement the drone-based vehicle testing method as described in any one of claims 10 to 13 when executing the computer program.
15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the drone-based vehicle testing method according to any one of claims 10 to 13.
