Vehicle running trajectory determination method, device, computer device and storage medium

Abstract

The present application relates to a method, apparatus, computer device, and storage medium for determining a vehicle running trajectory. The method includes: obtaining three-dimensional point cloud data and a plurality of image data of a target vehicle; performing feature matching on the three-dimensional point cloud data and the plurality of image data to obtain a matching result, and determining geometric parameter data of the target vehicle according to the matching result; obtaining outer contour data of the target vehicle according to the geometric parameter data; and determining the running trajectory of the target vehicle according to the outer contour data. By using this method, the accuracy of the geometric parameter data of the target vehicle can be improved, the accuracy of determining the actual running trajectory of the target vehicle can be ensured, and further the accuracy of the passing performance test result of the target vehicle can be improved.

Description

Technical Field

[0001] This application relates to the technical field of vehicle whole-vehicle detection, and particularly to a method and device for determining a vehicle running trajectory, a computer device, and a storage medium. Background Art

[0002] After a vehicle is assembled, it is necessary to conduct passability tests on various vehicle performances. Both the comprehensive performance detection line and the road simulation test are key technical means for performance testing, and are important guarantees for ensuring that the vehicle's various performances meet relevant standards and enterprise process specifications when leaving the factory.

[0003] Currently, the test items of the vehicle comprehensive performance detection line include non-contact four-wheel alignment, advanced driving assistance system (ADAS) calibration, 360-degree surround view system calibration, braking and antilock brake system (ABS) detection, high beam luminous intensity and high and low beam offset of headlamps, lateral slip amount of steering wheels, steering angle detection, external dimension measurement, etc., more than 10 items in total. Due to the large number of items, and in order to improve the detection rhythm and efficiency, generally, the detection line adopts a distributed measurement and control method, that is, these detection items are distributed in the detection workshop in the form of multiple workstations. The current specific implementation method of the detection items is that a vehicle driver drives the vehicle and, under the guidance of the driver assistant instrument, proceeds step by step according to the operation prompts set in the detection program. During the implementation of the above vehicle comprehensive performance detection line test items, it is necessary for the vehicle to maintain centering as much as possible, that is, the driving center line of the vehicle must be parallel or coincident with the center line of the detection line body, and as perpendicular as possible to the detection equipment (such as the running guide rail of the headlamp, the braking table body, the ADAS calibration frame, etc.). If the driving trajectory line of the vehicle does not meet the above requirements, that is, the driving center line of the vehicle deviates from the center line of the detection line body, on the one hand, it will lead to low accuracy, poor repeatability, and misjudgment of the passability test results, and on the other hand, it will also cause sampling, detection, and test failures due to excessive vehicle position deviation, resulting in repeated tests or vehicle return to the line, seriously affecting production efficiency.

[0004] The road simulation test is mainly carried out on a straight section. After the driver adjusts the vehicle steering wheel to be horizontal, the vehicle drives at a high speed along the ground markings. By observing the relationship between the vehicle and the ground markings with the eyes, and at the same time observing whether there is any left or right deviation of the steering wheel, the driver judges whether the vehicle has a tendency to run off course. It entirely depends on personal subjective experience, and there are problems such as "defective vehicles leaving the factory" for vehicles that run off course, and there are also certain potential risks.

[0005] Therefore, during the passability test of a vehicle, monitoring the driving trajectory of the vehicle and measuring and analyzing the actual movement trajectory of the vehicle can ensure the accuracy of the vehicle posture during the detection process, and further ensure the accuracy of the vehicle passability test.

[0006] In the traditional technology, satellite positioning is adopted, and the motion trajectory of the target vehicle is obtained by using an algorithm for fitting.

[0007] However, when performing passability tests on various performances of the target vehicle, due to the signal blind area of satellite positioning in indoor occlusion environments, and the irregular left and right oblique movements of the target vehicle may cause unstable reception of positioning signals, resulting in a low positioning accuracy of the target vehicle, the accuracy of the actual running trajectory of the target vehicle cannot be ensured, and further the accuracy of the passability test results of the target vehicle is reduced. Summary of the Invention

[0008] Based on this, in view of the above technical problems, it is necessary to provide a vehicle running trajectory determination method, device, computer device, and storage medium that can ensure the accuracy of the actual running trajectory of the target vehicle, and further improve the accuracy of the passability test results of the target vehicle.

[0009] In a first aspect, the present application provides a vehicle running trajectory determination method, including:

[0010] Obtain the three-dimensional point cloud data and multiple image data of the target vehicle;

[0011] Perform feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result;

[0012] Obtain the outer contour data of the target vehicle according to the geometric parameter data;

[0013] Determine the motion trajectory of the target vehicle according to the outer contour data.

[0014] In one embodiment, performing feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determining the geometric parameter data of the target vehicle according to the matching result includes:

[0015] Perform feature matching on the three-dimensional point cloud data and multiple image data to obtain matching result data; the matching result data includes multiple pose data of the target vehicle;

[0016] Fuse the three-dimensional point cloud data and multiple image data according to the matching result data and the fusion model to obtain the fused data of the target vehicle;

[0017] Extract the geometric parameter data of the target vehicle from the fused data.

[0018] In one embodiment, before performing feature matching on the three-dimensional point cloud data and multiple image data, it further includes:

[0019] Extract the three-dimensional feature data of the target vehicle from the three-dimensional point cloud data, and extract the two-dimensional feature data of the target vehicle from multiple image data;

[0020] Project the three-dimensional point cloud data of the target vehicle onto multiple image data;

[0021] Match the three-dimensional point cloud data with multiple image data according to the three-dimensional feature data and two-dimensional feature data of the target vehicle.

[0022] In one embodiment, extracting the three-dimensional feature data of the target vehicle from the three-dimensional point cloud data and extracting the two-dimensional feature data of the target vehicle from multiple image data includes:

[0023] Convert the three-dimensional point cloud data of the target vehicle into voxel grid data, and extract the three-dimensional feature data of the target vehicle from the voxel grid data;

[0024] Use a feature extraction algorithm to extract the key points and descriptors of the target vehicle from each image data to obtain the two-dimensional feature data of the target vehicle.

[0025] In one embodiment, obtaining the three-dimensional point cloud data and multiple image data of the target vehicle includes:

[0026] Obtain the original point cloud data of the target vehicle transmitted by the radar, and capture the original image data of the target vehicle transmitted by multiple vision sensors;

[0027] Adjust the original point cloud data and the original image data according to the timestamps of the original point cloud data and the original image data to obtain the time-synchronized original point cloud data and original image data;

[0028] Screen the time-synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data.

[0029] In one embodiment, after screening the time-synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data, it further includes:

[0030] Preprocess the screened original point cloud data and original image data to obtain the three-dimensional point cloud data of the target vehicle and multiple image data.

[0031] In one embodiment, the method further includes:

[0032] Perform a preliminary inspection on the geometric parameter data of the target vehicle to obtain the geometric parameter data after the preliminary inspection;

[0033] Obtain the historical parameter data of the target vehicle, and correct the geometric parameter data after the initial inspection according to the historical parameter data to obtain the corrected geometric parameter data;

[0034] Obtain the outline data of the target vehicle according to the corrected geometric parameter data;

[0035] Obtain the outline center point of the target vehicle according to the outline data, and determine the movement trajectory of the target vehicle according to the outline center point.

[0036] In a second aspect, the present application also provides a vehicle movement trajectory determination device, including:

[0037] A data acquisition module for acquiring the three-dimensional point cloud data and multiple image data of the target vehicle;

[0038] A fusion module for performing feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determining the geometric parameter data of the target vehicle according to the matching result;

[0039] An outline determination module for obtaining the outline data of the target vehicle according to the geometric parameter data;

[0040] A trajectory determination module for determining the movement trajectory of the target vehicle according to the outline data.

[0041] In a third aspect, the present application also provides a computer device, including a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0042] Acquire the three-dimensional point cloud data and multiple image data of the target vehicle;

[0043] Perform feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result;

[0044] Obtain the outline data of the target vehicle according to the geometric parameter data;

[0045] Determine the movement trajectory of the target vehicle according to the outline data.

[0046] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0047] Acquire the three-dimensional point cloud data and multiple image data of the target vehicle;

[0048] Perform feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result;

[0049] Based on the geometric parameter data, obtain the outline data of the target vehicle;

[0050] Based on the outline data, determine the motion trajectory of the target vehicle.

[0051] In a fifth aspect, the present application also provides a computer program product, including a computer program, which when executed by a processor implements the following steps:

[0052] Obtain the three-dimensional point cloud data and multiple image data of the target vehicle;

[0053] Perform feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result;

[0054] Based on the geometric parameter data, obtain the outline data of the target vehicle;

[0055] Based on the outline data, determine the motion trajectory of the target vehicle.

[0056] The above vehicle motion trajectory determination method, device, computer device, storage medium, and computer program product obtain the three-dimensional point cloud data and multiple image data of the target vehicle, and can obtain the relevant data of the target vehicle from different dimensions and angles, improving the reliability of the relevant data of the target vehicle; perform feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result, which can improve the accuracy of the geometric parameter data; based on the geometric parameter data, obtain the outline data of the target vehicle; based on the outline data, determine the motion trajectory of the target vehicle, which can ensure the accuracy of the determination of the actual motion trajectory of the target vehicle, and further improve the accuracy of the target vehicle passability test results. Description of the Drawings

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0058] Figure 1 It is an application environment diagram of the vehicle motion trajectory determination method in an embodiment;

[0059] Figure 2 It is a flowchart of the vehicle motion trajectory determination method in an embodiment;

[0060] Figure 3 It is a flowchart of the step of determining the geometric parameter data of the target vehicle in an embodiment;

[0061] Figure 4 It is a structural block diagram of a vehicle running trajectory determination device in an embodiment;

[0062] Figure 5 It is an internal structure diagram of a computer device in an embodiment. Specific embodiments

[0063] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0064] The vehicle running trajectory determination method provided by the embodiments of the present application can be applied to, for example, Figure 1 the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or can be placed in the cloud or other network servers. The terminal 102 sends a vehicle running trajectory determination request to the server 104, and the server 104 receives the vehicle running trajectory determination request, obtains the three-dimensional point cloud data and multiple image data of the target vehicle; performs feature matching on the three-dimensional point cloud data and the multiple image data to obtain a matching result, and determines the geometric parameter data of the target vehicle according to the matching result; determines the movement trajectory of the target vehicle according to the geometric parameter data. Among them, the terminal 102 can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0065] In an exemplary embodiment, as Figure 2 shown, a vehicle running trajectory determination method is provided. Taking the method applied to Figure 1 the server in as an example, it includes the following steps 202 to 208. Among them:

[0066] Step 202, obtain the three-dimensional point cloud data and multiple image data of the target vehicle.

[0067] Among them, the target vehicle refers to the vehicle that needs to conduct various performance passability tests. The three-dimensional point cloud data is a form of data obtained by a lidar or other sensors. The three-dimensional point cloud data of the target vehicle can represent the spatial structure of the target vehicle and its surrounding environment. The multiple image data of the target vehicle refers to the two-dimensional image data related to the target vehicle captured from different perspectives or different time points. The multiple two-dimensional image data includes the two-dimensional visual information of the target vehicle and its surrounding environment. The multiple image data can be obtained by multiple visual sensors or multiple cameras.

[0068] Exemplarily, the three-dimensional point cloud data around the target vehicle is obtained by a lidar, and the multiple two-dimensional image data of the target vehicle and its surrounding environment are captured by multiple visual sensors. The three-dimensional point cloud data obtained by the lidar can provide accurate depth and spatial information of the target vehicle, while the multiple two-dimensional image data captured by the visual sensors can capture more details and texture features related to the target vehicle. Therefore, based on the three-dimensional point cloud data obtained by the lidar and the multiple two-dimensional image data captured by the multiple visual sensors, rich information about the target vehicle and its surrounding environment can be obtained from different dimensions and angles of the target vehicle, thereby improving the reliability of the data related to the target vehicle.

[0069] Step 204: Perform feature matching on the three-dimensional point cloud data and the multiple image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result.

[0070] Among them, performing feature matching on the three-dimensional point cloud data and the multiple image data means that after determining the corresponding relationship between the three-dimensional point cloud data and the multiple image data, the spatial points in the three-dimensional point cloud data are docked with the pixel points in the image data, so that the information extracted from the image data can be combined with the point cloud in the three-dimensional data. Exemplarily, the matching result may include the position and attitude data of the target vehicle, etc.

[0071] Optionally, after extracting the key feature points and feature surface data related to the target vehicle from the three-dimensional point cloud data and extracting the feature data related to the target vehicle from each image data, the three-dimensional point cloud data is projected onto each image data, so that the key feature points and feature surface data related to the target vehicle extracted from the three-dimensional point cloud data are matched with the feature data related to the target vehicle extracted from each image data, and the matched feature points and feature surface data are obtained. The three-dimensional coordinate information of the target vehicle is extracted from the three-dimensional point cloud data obtained from the lidar, the corresponding relationship of these three-dimensional points in the two-dimensional image is determined by using the matched feature points and feature surface data, the two-dimensional projection information of these feature points is aligned with the three-dimensional coordinate data, and the geometric parameters such as the size, shape and posture of the vehicle are calculated by combining geometric algorithms, so as to obtain the geometric parameter data of the target vehicle, ensuring the accurate registration between the three-dimensional point cloud data and multiple image data, being able to obtain the feature information of the target vehicle more comprehensively and accurately, and further obtaining more accurate geometric parameter data.

[0072] Step 206, according to the geometric parameter data, obtain the outer contour data of the target vehicle.

[0073] Among them, the outer contour data of the target vehicle includes data such as the boundary shape and body contour line of the target vehicle.

[0074] Optionally, the geometric parameter data of the target vehicle may include basic information such as the length, width and height of the target vehicle. The body contour line of the target vehicle is constructed according to the geometric parameter data of the target vehicle. Among them, the body contour line of the target vehicle may be the outer contour of a rectangular frame, and this outer contour of the rectangular frame can effectively define the body range of the target vehicle, is not affected by irregular movements such as the left tilt and right tilt of the target vehicle, and stably represents the position and range of the target vehicle.

[0075] Step 208, according to the outer contour data, determine the movement trajectory of the target vehicle.

[0076] According to the outer contour data of the target vehicle, obtain the center point coordinates of the outer contour of the target vehicle, and fit and connect the center point coordinates of the outer contour of the target vehicle to obtain the movement trajectory of the target vehicle.

[0077] Exemplarily, for the outer contour data of the target vehicle in each frame, calculate the center point of the outer contour of the target vehicle in this frame. In continuous time frames, track the position change of the center point of the outer contour of the target vehicle frame by frame, and fit and connect the center points of the outer contour of the target vehicle to obtain the real-time movement trajectory of the target vehicle. Since the center point of the outer contour of the target vehicle is relatively stable and is not affected by the local morphological changes caused by the left and right tilting of the target vehicle, by fitting and connecting these center points of the outer contour of the target vehicle, and then forming the real-time movement trajectory of the target vehicle, it can more accurately reflect the overall movement path of the vehicle, thus effectively solving the problem that the trajectory accuracy is affected by the left and right tilting of the target vehicle.

[0078] In the above vehicle running trajectory determination method, obtaining the three-dimensional point cloud data and multiple image data of the target vehicle can obtain rich information about the target vehicle and its surrounding environment from different dimensions and angles of the target vehicle, thereby improving the reliability of the data related to the target vehicle; performing feature matching on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determining the geometric parameter data of the target vehicle according to the matching result can obtain the feature information of the target vehicle more comprehensively and accurately, thereby obtaining more accurate geometric parameter data; obtaining the outline data of the target vehicle according to the geometric parameter data can stably represent the position and range of the target vehicle; determining the running trajectory of the target vehicle according to the outline data can effectively solve the problem that the trajectory accuracy is affected by the left and right skew movement of the target vehicle, ensure the accuracy of determining the actual running trajectory of the target vehicle, and thereby improve the accuracy of the passing test result of the target vehicle.

[0079] In an exemplary embodiment, obtaining the three-dimensional point cloud data and multiple image data of the target vehicle includes: obtaining the original point cloud data of the target vehicle transmitted by a radar and capturing the original image data of the target vehicle transmitted by multiple vision sensors; adjusting the original point cloud data and the original image data according to the timestamps of the original point cloud data and the original image data to obtain the time-synchronized original point cloud data and original image data; screening the time-synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data.

[0080] Exemplarily, the original three-dimensional point cloud data of the target vehicle is obtained by a lidar sensor. These point cloud data include the point coordinate information of the target vehicle in three-dimensional space obtained from different angles. At the same time, multiple vision sensors are used to capture the original image data of the target vehicle from different perspectives. These image data contain the two-dimensional visual information and appearance feature data of the target vehicle. The original point cloud data and the original image data are adjusted by using the timestamps of the lidar sensor and multiple vision sensors to ensure the time synchronization of the original point cloud data and the original image data. By aligning the data captured at different time points, the time-synchronized original point cloud data and original image data can be obtained. Screening these time-synchronized original point cloud data and original image data. Among them, screening the time-synchronized original point cloud data and original image data includes applying a data preprocessing algorithm to filter out irrelevant parts and outliers from the time-synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data. The quality and accuracy of the screened original point cloud data and original image data can be ensured.

[0081] In this embodiment, time synchronization processing is performed on the original point cloud data of the target vehicle transmitted by the radar and the original image data of the target vehicle transmitted by multiple vision sensors, and the synchronized original point cloud data and original image data are screened, which can ensure the quality and accuracy of the screened original point cloud data and original image data.

[0082] In an exemplary embodiment, after screening the synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data, the following steps are further included: preprocessing the screened original point cloud data and original image data to obtain the three-dimensional point cloud data of the target vehicle and multiple image data.

[0083] Optionally, preprocessing the screened point cloud data and image data may include segmenting the screened original point cloud data, removing outliers, filling in missing data, and performing denoising, image enhancement, and feature extraction on the screened original image data, etc., to obtain the three-dimensional point cloud data of the target vehicle and multiple image data.

[0084] In this embodiment, preprocessing the screened original point cloud data and original image data can improve the reliability of the data related to the target vehicle.

[0085] In an exemplary embodiment, as Figure 3 shown, feature matching is performed on the three-dimensional point cloud data and multiple image data to obtain a matching result, and determining the geometric parameter data of the target vehicle according to the matching result includes steps 302 to 306. Among them:

[0086] Step 302, perform feature matching on the three-dimensional point cloud data and multiple image data to obtain matching result data; the matching result data includes multiple pose data of the target vehicle.

[0087] Step 304, fuse the three-dimensional point cloud data and multiple image data according to the matching result data and the fusion model to obtain the fused data of the target vehicle.

[0088] Step 306, extract the geometric parameter data of the target vehicle from the fused data.

[0089] Exemplarily, a feature matching algorithm is used to extract and match key feature points of three-dimensional point cloud data and multiple image data. The key feature points may include data such as vehicle body contours, edges, and corner points. Matching result data is generated, and the matching result data includes multiple pose data of the target vehicle at different viewpoints and time points, and these pose data describe the spatial position, direction, and pose changes of the target vehicle. Using the matching result data and a preset fusion model, data fusion of the three-dimensional point cloud data and the image data is performed. Among them, the fusion model can be obtained by training a deep learning model. The fusion model synthesizes data such as position, pose, direction, geometric features, and visual information in the three-dimensional point cloud data and the image data to generate fused data of the target vehicle, and the fused data provides more complete and consistent relevant data of the target vehicle. Geometric parameter data of the target vehicle is extracted from the fused data, and the geometric parameter data may include data such as the dimensions (such as length, width, and height), shape, wheel positions, vehicle body edges, and other geometric features of the target vehicle.

[0090] In the previous exemplary embodiment, before performing feature matching on the three-dimensional point cloud data and multiple image data, it further includes: extracting three-dimensional feature data of the target vehicle from the three-dimensional point cloud data, and extracting two-dimensional feature data of the target vehicle from the multiple image data; projecting the three-dimensional point cloud data of the target vehicle onto the multiple image data; and performing feature matching on the three-dimensional point cloud data and the multiple image data according to the three-dimensional feature data and the two-dimensional feature data of the target vehicle.

[0091] Exemplarily, the three-dimensional feature data may include data such as the geometric shape, vehicle body edge, and wheel position of the target vehicle, and the two-dimensional feature data may include the vehicle body contour of the target vehicle and other visual feature data. Projecting the three-dimensional point cloud data of the target vehicle onto multiple image data to visualize the three-dimensional geometric information of the target vehicle on the two-dimensional image data. Using the feature matching algorithm, according to the extracted three-dimensional feature data and two-dimensional feature data, feature matching is performed on the three-dimensional point cloud data and the multiple image data to identify and match the corresponding feature points of the vehicle at different viewpoints.

[0092] In this embodiment, by extracting features from the three-dimensional point cloud data and multiple image data, performing feature matching and data fusion, the fused data of the target vehicle can be identified and obtained more comprehensively and accurately, and the vehicle geometric parameters extracted from the fused data can improve the accuracy of the geometric parameter data of the target vehicle.

[0093] In the previous exemplary embodiment, extracting the three-dimensional feature data of the target vehicle from the three-dimensional point cloud data and extracting the two-dimensional feature data of the target vehicle from multiple image data includes: converting the three-dimensional point cloud data of the target vehicle into voxel grid data, and extracting the three-dimensional feature data of the target vehicle from the voxel grid data; using a feature extraction algorithm to extract the key points and descriptors of the target vehicle from each image data to obtain the two-dimensional feature data of the target vehicle.

[0094] Exemplarily, voxel grid data is a representation method that divides three-dimensional space into uniform volume units (i.e., voxels). Through voxelization processing, the three-dimensional point cloud data is converted into voxel grid data. In the voxel grid data, a three-dimensional feature extraction algorithm is used to extract the three-dimensional feature data of the target vehicle. Optionally, the three-dimensional feature extraction algorithm can adopt voxelized surface feature extraction (such as surface normal, curvature, etc.), three-dimensional shape descriptors, etc. From multiple image data, a feature extraction algorithm is used to extract the key points and descriptors of the target vehicle. Among them, the feature extraction algorithm includes SIFT (Scale-Invariant Feature Transform), SURF (Speeded-Up Robust Features), or ORB (Oriented FAST and Rotated BRIEF).

[0095] In this embodiment, converting the three-dimensional point cloud data into voxel grid data, extracting three-dimensional feature data therefrom, and using a feature extraction algorithm to extract two-dimensional feature data from the image data can make the feature matching between the three-dimensional point cloud data and multiple image data more accurate, and further make the geometric parameter data of the target vehicle more comprehensive.

[0096] In an exemplary embodiment, the method further includes: initially checking the geometric parameter data of the target vehicle to obtain the initially checked geometric parameter data; obtaining the historical parameter data of the target vehicle, and correcting the initially checked geometric parameter data according to the historical parameter data to obtain the corrected geometric parameter data; obtaining the outer contour data of the target vehicle according to the corrected geometric parameter data; obtaining the outer contour center point of the target vehicle according to the outer contour data, and determining the motion trajectory of the target vehicle according to the outer contour center point.

[0097] Exemplarily, initially checking the geometric parameter data of the target vehicle to obtain the initially checked geometric parameter data can ensure the integrity and accuracy of the geometric parameter data of the target vehicle. Obtaining the historical parameter data of the target vehicle, where the historical parameter data of the target vehicle can come from the vehicle model data in the database. Using the historical parameter data to correct the initially checked geometric parameter data, and the correction process can include comparing the initially checked geometric parameter data with the corresponding historical parameter data, adjusting and correcting the errors or deviations in the initially checked geometric parameter data to obtain more accurate and consistent corrected geometric parameter data.

[0098] Based on the corrected geometric parameter data, the outline data of the target vehicle is further generated. According to the outline data, the outline center point of the target vehicle is calculated, and the outline center point can be the position data of the geometric center of the vehicle. By tracking the change of the outline center point in consecutive time frames, the motion trajectory of the target vehicle is obtained.

[0099] In this embodiment, the initial inspection of the geometric parameter data of the target vehicle can ensure the integrity and accuracy of the geometric parameter data of the target vehicle. According to the outline data, determining the motion trajectory of the target vehicle can effectively solve the problem that the trajectory accuracy is affected by the left and right skew movement of the target vehicle, ensure the accuracy of the actual operation trajectory determination of the target vehicle, and further improve the accuracy of the passing test result of the target vehicle.

[0100] In another embodiment, a method for determining a vehicle operation trajectory is provided, and the method includes:

[0101] Obtain the original point cloud data of the target vehicle transmitted by the radar and capture the original image data of the target vehicle transmitted by multiple vision sensors; according to the timestamps of the original point cloud data and the original image data, adjust the original point cloud data and the original image data to obtain the time-synchronized original point cloud data and original image data; screen the time-synchronized original point cloud data and original image data to obtain the screened original point cloud data and original image data; preprocess the screened original point cloud data and original image data to obtain the three-dimensional point cloud data and multiple image data of the target vehicle.

[0102] Convert the three-dimensional point cloud data of the target vehicle into voxel grid data, and extract the three-dimensional feature data of the target vehicle from the voxel grid data; use a feature extraction algorithm to extract the key points and descriptors of the target vehicle from each image data to obtain the two-dimensional feature data of the target vehicle.

[0103] Project the three-dimensional point cloud data of the target vehicle onto multiple image data; according to the three-dimensional feature data and two-dimensional feature data of the target vehicle, perform feature matching on the three-dimensional point cloud data and multiple image data to obtain matching result data; the matching result data includes multiple pose data of the target vehicle.

[0104] According to the matching result data and the fusion model, fuse the three-dimensional point cloud data and multiple image data to obtain the fused data of the target vehicle; extract the geometric parameter data of the target vehicle from the fused data.

[0105] Conduct a preliminary inspection on the geometric parameter data of the target vehicle to obtain the geometric parameter data after the preliminary inspection; acquire the historical parameter data of the target vehicle, and correct the geometric parameter data after the preliminary inspection according to the historical parameter data to obtain the corrected geometric parameter data; obtain the outline data of the target vehicle based on the corrected geometric parameter data; obtain the outline center point of the target vehicle according to the outline data, and determine the movement trajectory of the target vehicle based on the outline center point

[0106] In this embodiment, time synchronization processing is performed on the original point cloud data of the target vehicle transmitted by the radar and the original image data of the target vehicle transmitted by multiple vision sensors, and the time-synchronized original point cloud data and original image data are screened, and the screened original point cloud data and original image data are preprocessed, which can ensure the quality and accuracy of the original point cloud data and original image data and improve the reliability of the data related to the target vehicle. By extracting features from the three-dimensional point cloud data and multiple image data, and performing feature matching and data fusion, the fused data of the target vehicle can be identified and obtained more comprehensively and accurately. The vehicle geometric parameters extracted from the fused data can improve the accuracy of the geometric parameter data of the target vehicle. Conducting a preliminary inspection on the geometric parameter data of the target vehicle can ensure the integrity and accuracy of the geometric parameter data of the target vehicle. Determining the movement trajectory of the target vehicle based on the outline data can effectively solve the problem that the trajectory accuracy is affected by the left-right skew movement of the target vehicle, ensure the accuracy of determining the actual running trajectory of the target vehicle, and further improve the accuracy of the passing test results of the target vehicle.

[0107] Through the above vehicle running trajectory determination method, the actual parameters of the target vehicle can be obtained, the actual running trajectory of the target vehicle can be monitored, deviation warning can be carried out to ensure vehicle centering, and it can also be used in vehicle passing tests. For example, after receiving the geometric parameter data sent from the production management system, a theoretical vehicle profile is generated, and then the actual production environment vehicle profile is obtained through the lidar system. After verification, it can be found whether there are production passing risks and problems for a certain model, including but not limited to length, height, width, and chassis clearance, and it can be confirmed whether there is interference between the vehicle profile and the site. After identifying the risk sources, they can be quickly processed. By monitoring the actual running trajectory of the vehicle, the driver's driving direction can be accurately guided, and it can be confirmed whether there are problems in the actual test process of the vehicle detection line.

[0108] It should be understood that although the steps in the flowcharts involved in the above embodiments are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, there is no strict order limit for the execution of these steps, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or steps or stages in other steps.

[0109] Based on the same inventive concept, an embodiment of the present application further provides a vehicle running trajectory determination device for implementing the vehicle running trajectory determination method described above. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the vehicle running trajectory determination device provided below can refer to the limitations on the vehicle running trajectory determination method in the above text, and will not be repeated here.

[0110] In an exemplary embodiment, as Figure 4 shown, a vehicle running trajectory determination device is provided, including: a data acquisition module 402, a fusion module 404, an outline determination module 406, and a trajectory determination module 408, where:

[0111] The data acquisition module 402 is configured to acquire three-dimensional point cloud data and a plurality of image data of a target vehicle.

[0112] The fusion module 404 is configured to perform feature matching on the three-dimensional point cloud data and the plurality of image data to obtain a matching result, and determine geometric parameter data of the target vehicle according to the matching result.

[0113] The outline determination module 406 is configured to obtain outline data of the target vehicle according to the geometric parameter data.

[0114] The trajectory determination module 408 is configured to determine the motion trajectory of the target vehicle according to the outline data.

[0115] In an exemplary embodiment, the data acquisition module 402 includes a lidar interface unit, a vision sensor interface unit, a data buffer unit, a synchronization unit, and a data processing unit. The lidar interface unit is responsible for communicating with the lidar sensor and receiving the raw point cloud data or distance data sent by it. The lidar interface unit is also used to process the lidar-specific communication protocol and data format. The vision sensor interface unit is responsible for communicating with a vision sensor (such as a camera) and receiving the raw image data captured by it. The vision sensor interface unit is also used to process the synchronization, decompression, and format conversion of the image data. The data buffer unit is responsible for temporarily storing the raw point cloud data and raw image data received from the lidar sensor and the vision sensor. The data buffer unit is also used to smooth the difference between the data transmission rate and the processing speed. When the data processing unit is ready to process the data, the data buffer unit sends the raw point cloud data and raw image data to the processing unit. The synchronization unit is responsible for ensuring that the data received from the lidar sensor and the vision sensor are synchronized in time. The synchronization unit is used to process the timestamps, frame rates, etc. of the lidar sensor and the vision sensor. The data processing unit is used to verify the integrity of the raw point cloud data and raw image data, and perform processing such as filtering noise or outliers.

[0116] The lidar interface unit receives the raw point cloud data from the lidar sensor and stores the raw point cloud data in the data buffer unit. The vision sensor interface unit receives the raw image data from the vision sensor and stores the raw image data in the data buffer unit. The synchronization unit reads the raw point cloud data and raw image data from the data buffer unit, and adjusts or marks the timestamps as needed to ensure the synchronization of the raw point cloud data and the raw image data. The synchronization unit sends the synchronized raw point cloud data and raw image data to the data processing unit for preliminary processing. After the data processing unit completes the preliminary data check and screening of the raw point cloud data and the raw image data, it sends the processed raw point cloud data and raw image data to the preprocessing module.

[0117] In an exemplary embodiment, the vehicle running trajectory determination device further includes:

[0118] The preprocessing module includes a data reception unit, a denoising unit, a compression unit, and a data output unit. The data reception unit is responsible for receiving the processed original point cloud data and original image data, ensuring the integrity and accuracy of the processed original point cloud data and original image data. The denoising unit is responsible for denoising the received processed original point cloud data and original image data, removing noise and outliers, and obtaining the denoised original point cloud data and original image data. The denoising algorithm can include statistical filtering, median filtering, etc. The compression unit is responsible for compressing the denoised original point cloud data and original image data to reduce the overhead of data transmission and storage. The compression algorithm can include point cloud compression, image compression, etc. After the above preprocessing, the three-dimensional point cloud data and multiple image data of the target vehicle are obtained. The data output unit is responsible for sending the three-dimensional point cloud data and multiple image data of the target vehicle to the fusion module. At the same time, the three-dimensional point cloud data and multiple image data of the target vehicle are stored in the intelligent cache module.

[0119] The data reception unit receives the processed original point cloud data and original image data, and preprocesses the processed original point cloud data and original image data. The preprocessing process includes sending the processed original point cloud data and original image data to the denoising unit for denoising. The denoising unit sends the denoised original point cloud data and original image data to the compression unit for compression operation to obtain the three-dimensional point cloud data and multiple image data of the target vehicle. The data output unit then sends the three-dimensional point cloud data and multiple image data of the target vehicle to the fusion module and the intelligent cache module.

[0120] In an exemplary embodiment, the vehicle running trajectory determination device further includes:

[0121] The intelligent cache module includes a cache storage unit, a cache management unit, a cache interface unit, and a cache policy unit. The cache storage unit is responsible for actually storing data, which can be memory, hard disk, or other storage media. The cache storage unit supports fast data read and write operations. The cache management unit is responsible for managing the data in the cache storage unit, including adding, retrieving, replacing, and updating data, and deciding which data should be retained or replaced according to the cache policy. The cache interface unit is used to provide an interface for interacting with other modules (such as the fusion module, post-processing module). The cache interface unit receives requests from other modules, such as data retrieval or data update, and calls the cache management unit to perform corresponding operations. The cache policy unit is responsible for defining and maintaining the cache policy, such as LRU (Least Recently Used), LFU (Least Frequently Used), etc. The cache policy unit selects a suitable cache policy according to the access pattern of the data and the system requirements. Among them, the data can include the three-dimensional point cloud data and multiple image data of the target vehicle.

[0122] The cache interface unit receives data retrieval requests from the fusion module or the post-processing module. The cache interface unit forwards the requests to the cache management unit and returns the retrieved data to the requesting module. When new data needs to be stored, the cache management unit is responsible for writing the data into the cache storage unit and decides whether to replace old data according to the cache policy. The cache management unit queries the cache policy unit when needed to obtain the currently used cache policy. The cache management unit decides data replacement and update according to the cache policy. The preprocessing module sends the three-dimensional point cloud data and multiple image data of the target vehicle to the cache interface unit for storage.

[0123] In this embodiment, through the intelligent cache module, common data or calculation results are cached, and the calculation and transmission time are reduced by storing and reusing the data, thereby improving the real-time performance and response speed of the device to provide the required data in a timely and accurate manner during vehicle driving.

[0124] In an exemplary embodiment, the fusion module 404 includes a data acquisition unit, a data fusion processing unit, a cache access unit, and a fusion result output unit. The data acquisition unit is responsible for receiving the three-dimensional point cloud data and multiple image data of the target vehicle sent by the preprocessing module, ensuring the integrity and accuracy of the three-dimensional point cloud data and multiple image data of the target vehicle, and performing necessary data verification. The data fusion processing unit is responsible for performing multi-sensor data fusion using the trained model, accessing the data in the intelligent cache module to reduce repeated calculations, combining the three-dimensional point cloud data and multiple image data of the target vehicle, and extracting and fusing to obtain the geometric parameter data of the target vehicle. The cache access unit is responsible for interacting with the intelligent cache module to retrieve or update the three-dimensional point cloud data and multiple image data of the target vehicle in the cache. During the fusion process of the three-dimensional point cloud data and multiple image data of the target vehicle, historical data or intermediate results are quickly retrieved from the intelligent cache module as needed. The fusion result output unit is responsible for sending the geometric parameter data of the target vehicle obtained by the data fusion processing unit to the post-processing module and storing the fusion result in the intelligent cache module for subsequent use or analysis.

[0125] The data acquisition unit receives the three-dimensional point cloud data and multiple image data of the target vehicle from the preprocessing module and transfers them to the data fusion processing unit for fusion processing. The data fusion processing unit retrieves the three-dimensional point cloud data and multiple image data of the target vehicle from the intelligent cache module through the cache access unit when needed. The data fusion processing unit stores the intermediate result or the final fusion result in the intelligent cache module through the cache access unit. After the data fusion processing unit completes the data fusion, it sends the fusion result to the fusion result output unit.

[0126] In an exemplary embodiment, the vehicle running trajectory determination device further includes:

[0127] The post - processing module includes a fusion data receiving unit, a fusion data processing unit, an intelligent cache access unit, and a post - processing result output unit. The fusion data receiving unit is responsible for receiving the geometric parameter data of the target vehicle output by the fusion module, and performing preliminary verification on the received geometric parameter data of the target vehicle to ensure the integrity and accuracy of the geometric parameter data of the target vehicle. The fusion data processing unit is responsible for further processing and analyzing the received geometric parameter data of the target vehicle. The fusion data processing unit accesses the data in the intelligent cache module to obtain historical measurement results for comparison or analysis, and applies various algorithms and models to optimize, correct, or verify the geometric parameter data of the target vehicle. The intelligent cache access unit is responsible for interacting with the intelligent cache module to retrieve historical measurement data or other relevant information. The intelligent cache access unit provides a data retrieval function so that the fusion data processing unit can obtain the required historical data. The post - processing result output unit is responsible for sending the processed geometric parameter data of the target vehicle to the terminal for display on the user interface or for use by other systems, and storing the processed results in the intelligent cache module for subsequent analysis and query.

[0128] In an exemplary embodiment, the vehicle running trajectory determination device further includes:

[0129] The model management module includes a model storage unit, a model management unit, a model call interface unit, and a model update and maintenance unit. The model storage unit is responsible for storing the trained models, including model parameters, structures, and related metadata, providing an efficient storage solution to ensure the integrity and accessibility of the model data. The model management unit is responsible for managing the models in the model storage unit, including adding, deleting, modifying, and querying models, maintaining model version information to ensure that the system can use the latest model version, monitoring the performance and status of the models, and providing suggestions for model update and maintenance. The model call interface unit is responsible for providing interfaces for interacting with other modules (such as the fusion module), allowing them to call the models for data processing as needed, verifying the permissions and legality of the requesting module, and ensuring the secure use of the models. The model update and maintenance unit is responsible for the update and maintenance work of the models, including downloading new model versions, replacing old versions, performing necessary model verification and testing, monitoring the performance and stability of the models, and promptly discovering and fixing potential problems. Among them, the models can include fusion models.

[0130] The model management unit adds, deletes, modifies, and queries model data through the model storage unit. When other modules (such as the fusion module) need to call the model, they send requests to the model management unit through the model call interface unit. The model update and maintenance unit receives the update and maintenance instructions from the model management unit and performs corresponding operations. The model update and maintenance unit obtains the new version of the model data from an external source and stores it in the model storage unit.

[0131] In this embodiment, the trained models are stored and managed by the model management unit. When other modules need them, the models can be called for data processing, and the efficient call mechanism can reduce the generation time of model loading and initialization.

[0132] In an exemplary embodiment, the trajectory determination module 408 is further configured to perform a preliminary inspection on the geometric parameter data of the target vehicle to obtain the geometric parameter data after the preliminary inspection; obtain the historical parameter data of the target vehicle, and correct the geometric parameter data after the preliminary inspection according to the historical parameter data to obtain the corrected geometric parameter data; obtain the outline data of the target vehicle according to the corrected geometric parameter data; obtain the outline center point of the target vehicle according to the outline data, and determine the movement trajectory of the target vehicle according to the outline center point.

[0133] In this embodiment, through the trajectory determination module, trajectory prompts and monitoring can be carried out at workstations such as four-wheel alignment, driving assistance, lighting detection, and corner detection on the vehicle comprehensive performance detection line, ensuring the accuracy and precision of vehicle comprehensive performance detection and improving vehicle production efficiency; at the same time, it can also judge the passability problem of vehicle models through the constructed geometric parameter data of the vehicle.

[0134] Each module in the above vehicle running trajectory determination device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above respective modules.

[0135] In an exemplary embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 5As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store three-dimensional point cloud data, multiple image data, geometric parameter data, outline data, etc. of the target vehicle. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a method for determining a vehicle running trajectory.

[0136] Those skilled in the art can understand that Figure 5 the structure shown in the figure is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0137] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0138] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0139] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0140] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0141] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., and are not limited thereto. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., and are not limited thereto.

[0142] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0143] The above embodiments only represent several implementation manners of the present application. The description is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for determining a vehicle running trajectory, characterized in that, The method includes: Obtaining three-dimensional point cloud data and multiple image data of a target vehicle; the three-dimensional point cloud data represents the spatial structure of the target vehicle and its surrounding environment; the multiple image data refers to two-dimensional image data related to the target vehicle captured from different perspectives or at different time points, and the multiple two-dimensional image data includes two-dimensional visual information of the target vehicle and its surrounding environment; Extracting three-dimensional feature data of the target vehicle from the three-dimensional point cloud data, and extracting two-dimensional feature data of the target vehicle from the multiple image data; projecting the three-dimensional point cloud data of the target vehicle onto the multiple image data; according to the three-dimensional feature data and two-dimensional feature data of the target vehicle, performing feature matching on the three-dimensional point cloud data and the multiple image data to obtain a matching result, and determining geometric parameter data of the target vehicle according to the matching result, including: performing feature matching on the three-dimensional point cloud data and the multiple image data to obtain matching result data; the matching result data includes multiple pose data of the target vehicle; according to the matching result data and a fusion model, fusing the three-dimensional point cloud data and the multiple image data to obtain fused data of the target vehicle; extracting the geometric parameter data of the target vehicle from the fused data; Obtaining outer contour data of the target vehicle according to the geometric parameter data; Determining the motion trajectory of the target vehicle according to the outer contour data.

2. The method according to claim 1, characterized in that, The extracting three-dimensional feature data of the target vehicle from the three-dimensional point cloud data and extracting two-dimensional feature data of the target vehicle from the multiple image data includes: Converting the three-dimensional point cloud data of the target vehicle into voxel grid data, and extracting the three-dimensional feature data of the target vehicle from the voxel grid data; Using a feature extraction algorithm to extract key points and descriptors of the target vehicle from each of the image data to obtain two-dimensional feature data of the target vehicle.

3. The method according to claim 1, wherein The obtaining three-dimensional point cloud data and multiple image data of the target vehicle includes: Obtaining the original point cloud data of the target vehicle transmitted by a radar, and capturing the original image data of the target vehicle transmitted by multiple vision sensors; Adjusting the original point cloud data and the original image data according to the timestamps of the original point cloud data and the original image data to obtain time-synchronized original point cloud data and original image data; Screening the time-synchronized original point cloud data and original image data to obtain screened original point cloud data and original image data.

4. The method according to claim 3, characterized in that, After the screening the time-synchronized original point cloud data and original image data to obtain screened original point cloud data and original image data, it further includes: Performing preprocessing on the screened original point cloud data and original image data to obtain three-dimensional point cloud data and multiple image data of the target vehicle.

5. The method according to claim 1, wherein The method further includes: Performing preliminary inspection on the geometric parameter data of the target vehicle to obtain preliminarily inspected geometric parameter data; Obtaining historical parameter data of the target vehicle, and correcting the preliminarily inspected geometric parameter data according to the historical parameter data to obtain corrected geometric parameter data; Obtain the outline data of the target vehicle according to the corrected geometric parameter data; According to the outline data, obtain the center point of the outline of the target vehicle, and determine the motion trajectory of the target vehicle according to the center point of the outline.

6. A vehicle running trajectory determination device, characterized in that, The device includes: A data acquisition module, configured to acquire three-dimensional point cloud data and a plurality of image data of the target vehicle; the three-dimensional point cloud data represents the spatial structure of the target vehicle and its surrounding environment; the plurality of image data refers to two-dimensional image data related to the target vehicle captured from different perspectives or different time points, and the plurality of two-dimensional image data includes two-dimensional visual information of the target vehicle and its surrounding environment; A fusion module, configured to extract three-dimensional feature data of the target vehicle from the three-dimensional point cloud data, and extract two-dimensional feature data of the target vehicle from the plurality of image data; project the three-dimensional point cloud data of the target vehicle onto the plurality of image data; according to the three-dimensional feature data and two-dimensional feature data of the target vehicle, perform feature matching on the three-dimensional point cloud data and the plurality of image data to obtain a matching result, and determine the geometric parameter data of the target vehicle according to the matching result, including: performing feature matching on the three-dimensional point cloud data and the plurality of image data to obtain matching result data; the matching result data includes a plurality of pose data of the target vehicle; according to the matching result data and the fusion model, fuse the three-dimensional point cloud data and the plurality of image data to obtain the fused data of the target vehicle; extract the geometric parameter data of the target vehicle from the fused data; An outline determination module, configured to obtain the outline data of the target vehicle according to the geometric parameter data; A trajectory determination module, configured to determine the motion trajectory of the target vehicle according to the outline data.

7. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images