Positioning compensation method and device

Abstract

The invention discloses a positioning compensation method and device, and the method comprises the steps: collecting vehicle image data corresponding to a target vehicle, extracting image feature points in the vehicle image data, carrying out the feature point matching operation of all image feature points, and obtaining a feature matching result; based on a feature matching result, determining image depth information corresponding to the target vehicle; according to the image depth information, track prediction operation is executed on the target vehicle, and a vehicle motion track corresponding to the target vehicle is obtained; and based on the vehicle movement track and the current positioning information corresponding to the target vehicle, performing a positioning compensation operation on the target vehicle to obtain a positioning result corresponding to the target vehicle. Therefore, by implementing the method and the system, the positioning of the vehicle can be compensated intelligently, the positioning accuracy and reliability can be improved, and the positioning intelligence can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of positioning compensation, and in particular to a positioning compensation method and device. Background Art

[0002] Currently, the RTK (Real-Time Kinematic) positioning solution is a technology that uses the satellite navigation system (GNSS) to provide high-precision positioning. RTK can achieve centimeter-level positioning accuracy through differential correction signals between the base station and the mobile station. Its principle is that the mobile station receives signals from multiple satellites and eliminates errors through differential correction information provided by the base station, thereby achieving high-precision positioning results.

[0003] In vehicle positioning, RTK solutions are widely used, especially in the fields of autonomous driving, drones, agricultural automation, etc. The advantages of RTK positioning systems are their high accuracy and small positioning error, which can meet the needs of precise navigation. However, due to the strong dependence of RTK systems on external conditions, their stability and reliability in complex environments will be affected to a certain extent, resulting in positioning deviations or errors. It can be seen that it is particularly important to provide a new positioning compensation method to improve the accuracy and reliability of positioning. Summary of the invention

[0004] The present invention provides a positioning compensation method and device, which can intelligently compensate for the positioning of a vehicle, thereby helping to improve the positioning accuracy and reliability, and can also help to improve the intelligence of positioning.

[0005] In order to solve the above technical problems, the first aspect of the present invention discloses a positioning compensation method, the method comprising:

[0006] Collecting vehicle image data corresponding to the target vehicle, extracting image feature points in the vehicle image data, and performing a feature point matching operation on all the image feature points to obtain a feature matching result;

[0007] Based on the feature matching result, determining the image depth information corresponding to the target vehicle;

[0008] According to the image depth information, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle;

[0009] Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle.

[0010] As an optional implementation manner, in the first aspect of the present invention, after performing a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle and obtaining the positioning result corresponding to the target vehicle, the method further includes:

[0011] Acquiring motion data of the target vehicle, wherein the motion data of the target vehicle includes one or more of relative displacement data, relative direction data, motion speed data, and motion acceleration data of the target vehicle;

[0012] Generate relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle;

[0013] According to the relative motion information, a positioning update operation is performed on the positioning result corresponding to the target vehicle.

[0014] As an optional implementation, in the first aspect of the present invention, before performing a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle, the method further includes:

[0015] Acquire an image frame data set corresponding to the target vehicle, wherein the image frame data set includes a plurality of frames of image data corresponding to the target vehicle, and all of the frame image data are continuous image frames;

[0016] Performing a posture prediction operation on the image frame data set through a preset Kalman filter to obtain a posture prediction result corresponding to the image frame data set, and performing a data filtering operation on the posture prediction result to obtain a target posture result;

[0017] The step of performing a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle includes:

[0018] According to the image depth information and the target pose result, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle.

[0019] As an optional implementation, in the first aspect of the present invention, before determining the image depth information corresponding to the target vehicle based on the feature matching result, the method further includes:

[0020] Determine whether there are matching points to be eliminated in the feature matching results;

[0021] When it is determined that the to-be-removed matching point exists in the feature matching result, based on a predetermined target algorithm, a data removal operation is performed on the to-be-removed matching point in the feature matching result to update the feature matching result;

[0022] Analyze the updated feature matching result to obtain rotation matrix parameters and translation vector parameters corresponding to the feature matching result, and generate relative motion parameters based on the rotation matrix parameters and the translation vector parameters;

[0023] Wherein, determining the image depth information corresponding to the target vehicle based on the feature matching result includes:

[0024] Based on the feature matching result and the relative motion parameter, image depth information corresponding to the target vehicle is determined.

[0025] As an optional implementation manner, in the first aspect of the present invention, determining the image depth information corresponding to the target vehicle based on the feature matching result and the relative motion parameter, determining the image depth information corresponding to the target vehicle includes:

[0026] Based on the feature matching result, at least one target frame data group is determined from all the frame image data, wherein the target frame data group includes at least two frame image data, and the corresponding order of all the frame image data included in each target frame data group is an adjacent order;

[0027] Determine feature point change information corresponding to each target frame data group according to the relative motion parameters, and perform a coordinate estimation operation on the feature matching result based on the feature point change information corresponding to all the target frame data groups to obtain spatial coordinate information corresponding to the feature matching result, and determine image depth information corresponding to the target vehicle based on the spatial coordinate information;

[0028] The image depth information includes depth information corresponding to the current environment where the target vehicle is located.

[0029] As an optional implementation manner, in the first aspect of the present invention, performing a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle includes:

[0030] Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, determining the relative position change information corresponding to the target vehicle;

[0031] According to the relative posture change information, the relative displacement change information of the target vehicle and the relative path change information of the target vehicle are determined, and according to the relative displacement change information and the relative path change information, the positioning result corresponding to the target vehicle is determined.

[0032] As an optional implementation manner, in the first aspect of the present invention, determining the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information includes:

[0033] Acquire environmental information of the current environment where the target vehicle is located, and construct an environmental map corresponding to the current environment based on the environmental information;

[0034] Extracting environmental feature information from the environmental map, and determining position change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information;

[0035] A feature point matching operation is performed on the environmental feature information and the posture change feature information to obtain a posture feature matching result corresponding to the target vehicle, and a positioning result corresponding to the target vehicle is determined based on the posture feature matching result.

[0036] A second aspect of the present invention discloses a positioning compensation device, the device comprising:

[0037] An acquisition module, used for acquiring vehicle image data corresponding to a target vehicle;

[0038] An extraction module, used to extract image feature points from the vehicle image data;

[0039] A matching module, used to perform a feature point matching operation on all the image feature points to obtain a feature matching result;

[0040] A determination module, used to determine the image depth information corresponding to the target vehicle based on the feature matching result;

[0041] A prediction module, configured to perform a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle;

[0042] The positioning module is used to perform a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle.

[0043] As an optional implementation, in the second aspect of the present invention, an acquisition module is used to obtain motion data of the target vehicle after the positioning module performs a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain the positioning result corresponding to the target vehicle, wherein the motion data of the target vehicle includes one or more of relative displacement data, relative direction data, motion speed data, and motion acceleration data of the target vehicle;

[0044] A generating module, configured to generate relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle;

[0045] An updating module is used to perform a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information.

[0046] As an optional implementation, in the second aspect of the present invention, the acquisition module is further used to acquire an image frame data set corresponding to the target vehicle before the prediction module performs a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle, wherein the image frame data set includes a plurality of frame image data corresponding to the target vehicle, and all the frame image data are continuous image frames;

[0047] The prediction module is further used to perform a posture prediction operation on the image frame data set through a preset Kalman filter to obtain a posture prediction result corresponding to the image frame data set, and perform a data filtering operation on the posture prediction result to obtain a target posture result;

[0048] The prediction module performs a trajectory prediction operation on the target vehicle according to the image depth information, and obtains the vehicle motion trajectory corresponding to the target vehicle in a specific manner including:

[0049] According to the image depth information and the target pose result, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle.

[0050] As an optional implementation, in the second aspect of the present invention, the device further includes:

[0051] A judging module, used for judging whether there are matching points to be eliminated in the feature matching results before the determining module determines the image depth information corresponding to the target vehicle based on the feature matching results;

[0052] A removal module, configured to, when the judgment module judges that the matching point to be removed exists in the feature matching result, perform a data removal operation on the matching point to be removed in the feature matching result based on a predetermined target algorithm, so as to update the feature matching result;

[0053] The generating module is further used to analyze the updated feature matching result, obtain the rotation matrix parameters and the translation vector parameters corresponding to the feature matching result, and generate relative motion parameters based on the rotation matrix parameters and the translation vector parameters;

[0054] The specific manner in which the determination module determines the image depth information corresponding to the target vehicle based on the feature matching result includes:

[0055] Based on the feature matching result and the relative motion parameter, image depth information corresponding to the target vehicle is determined.

[0056] As an optional implementation, in the second aspect of the present invention, the determination module determines the image depth information corresponding to the target vehicle based on the feature matching result and the relative motion parameter, and the specific manner of determining the image depth information corresponding to the target vehicle includes:

[0057] Based on the feature matching result, at least one target frame data group is determined from all the frame image data, wherein the target frame data group includes at least two frame image data, and the corresponding order of all the frame image data included in each target frame data group is an adjacent order;

[0058] Determine feature point change information corresponding to each target frame data group according to the relative motion parameters, and perform a coordinate estimation operation on the feature matching result based on the feature point change information corresponding to all the target frame data groups to obtain spatial coordinate information corresponding to the feature matching result, and determine image depth information corresponding to the target vehicle based on the spatial coordinate information;

[0059] The image depth information includes depth information corresponding to the current environment where the target vehicle is located.

[0060] As an optional implementation, in the second aspect of the present invention, the positioning module performs a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, and the specific manner of obtaining the positioning result corresponding to the target vehicle includes:

[0061] Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, determining the relative position change information corresponding to the target vehicle;

[0062] According to the relative posture change information, the relative displacement change information of the target vehicle and the relative path change information of the target vehicle are determined, and according to the relative displacement change information and the relative path change information, the positioning result corresponding to the target vehicle is determined.

[0063] As an optional implementation, in the second aspect of the present invention, the specific manner in which the positioning module determines the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information includes:

[0064] Acquire environmental information of the current environment where the target vehicle is located, and construct an environmental map corresponding to the current environment based on the environmental information;

[0065] Extracting environmental feature information from the environmental map, and determining position change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information;

[0066] A feature point matching operation is performed on the environmental feature information and the posture change feature information to obtain a posture feature matching result corresponding to the target vehicle, and a positioning result corresponding to the target vehicle is determined based on the posture feature matching result.

[0067] The third aspect of the present invention discloses another positioning compensation device, the device comprising:

[0068] A memory storing executable program code;

[0069] a processor coupled to the memory;

[0070] The processor calls the executable program code stored in the memory to execute the positioning compensation method disclosed in the first aspect of the present invention.

[0071] The fourth aspect of the present invention discloses a computer storage medium, wherein the computer storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the positioning compensation method disclosed in the first aspect of the present invention.

[0072] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0073] In an embodiment of the present invention, vehicle image data corresponding to the target vehicle is collected, image feature points in the vehicle image data are extracted, and a feature point matching operation is performed on all image feature points to obtain a feature matching result; based on the feature matching result, the image depth information corresponding to the target vehicle is determined; according to the image depth information, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle; based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle. It can be seen that the implementation of the present invention can intelligently compensate for the positioning of the vehicle, which is beneficial to improving the positioning accuracy and reliability, and can also be beneficial to improving the intelligence of positioning. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0075] Figure 1 It is a flow chart of a positioning compensation method disclosed in an embodiment of the present invention;

[0076] Figure 2 It is a flow chart of another positioning compensation method disclosed in an embodiment of the present invention;

[0077] Figure 3 It is a structural schematic diagram of a positioning compensation device disclosed in an embodiment of the present invention;

[0078] Figure 4 It is a structural schematic diagram of another positioning compensation device disclosed in an embodiment of the present invention;

[0079] Figure 5 It is a structural schematic diagram of another positioning compensation device disclosed in an embodiment of the present invention. DETAILED DESCRIPTION

[0080] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0081] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or end including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units inherent to these processes, methods, products or ends.

[0082] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0083] The present invention discloses a positioning compensation method and device, which can intelligently compensate for the positioning of a vehicle, thereby improving the positioning accuracy and reliability, and also improving the intelligence of positioning. The following are detailed descriptions of each method and device.

[0084] Embodiment 1

[0085] See also Figure 1 , Figure 1 : is a flow chart of a positioning compensation method disclosed in an embodiment of the present invention. Figure 1 The positioning compensation method described can be applied to a positioning compensation device or to the vehicle itself, wherein the positioning compensation device can be integrated in a cloud server or a local server, which is not limited in the embodiments of the present invention. Figure 1 As shown, the positioning compensation method may include the following operations:

[0086] 101. Collect vehicle image data corresponding to the target vehicle, extract image feature points in the vehicle image data, and perform feature point matching operations on all image feature points to obtain feature matching results.

[0087] In an embodiment of the present invention, optionally, the vehicle image data corresponding to the target vehicle can be acquired by one or more of a monocular camera and a binocular camera; in this way, the vehicle image data of the target vehicle can be acquired at a low cost through a monocular camera, and the depth information can be directly acquired through the two lenses of a binocular camera, which can help improve the positioning accuracy; further, the vehicle image data corresponding to the target vehicle can also be acquired by one or more of a laser radar sensor and an ultrasonic sensor; wherein, the laser radar sensor (LiDAR, Light Detection and Ranging) is a non-contact detection and ranging technology, which accurately calculates the distance of an object by emitting laser pulses and measuring the reflected light; an ultrasonic sensor is a sensing device that uses ultrasonic signals to detect the position, distance or speed of an object.

[0088] In the embodiment of the present invention, further optionally, the vehicle image data of the target vehicle may be collected by collecting the image data of the target vehicle through a sensor pre-installed on the target vehicle.

[0089] In the embodiment of the present invention, optionally, the above-mentioned extraction of image feature points in the vehicle image data and performing a feature point matching operation on all image feature points to obtain a feature matching result may include:

[0090] Extracting image feature points from the vehicle image data by using a predetermined feature detection algorithm, and generating feature description information corresponding to each image feature point; wherein the predetermined feature detection algorithm may include an ORB detection algorithm;

[0091] By using a predetermined feature matching algorithm, feature point pairs are determined in the feature description information, and feature matching results are generated based on all feature point pairs; wherein the predetermined feature matching algorithm may include a FLANN matching algorithm, and the feature matching results may include at least all feature point pairs.

[0092] In the embodiment of the present invention, the optional ORB detection algorithm, namely Oriented FAST and Rotated BRIEF, is a feature detection and description algorithm widely used in the field of computer vision. It combines the advantages of the FAST corner detection algorithm and the BRIEF descriptor algorithm to achieve efficient and rotation-invariant feature extraction and matching; wherein the ORB algorithm consists of two parts: feature point extraction and feature point description. Among them, the feature point extraction part adopts the FAST (Features from Accelerated Segment Test) algorithm, and the feature point description part is based on the BRIEF (Binary Robust Independent Elementary Features) feature description algorithm. Therefore, the ORB algorithm can be regarded as a combination of FAST and BRIEF. FLANN (Fast Library for Approximate Nearest Neighbors) is a library for performing fast approximate nearest neighbor searches. It integrates a variety of advanced search algorithms and can automatically select the optimal algorithm according to the characteristics of the data, thereby achieving efficient search performance.

[0093] In an embodiment of the present invention, optionally, for example, ORB or other feature point detection algorithms are used to extract feature points in each frame of the image, and descriptors (local feature descriptions of feature points) are generated. Matching feature point pairs are found in adjacent image frames through nearest neighbor search-based methods such as FLANN (Fast Library for Approximate Nearest Neighbors), and feature matching results are generated based on the feature point pairs.

[0094] 102. Based on the feature matching result, determine the image depth information corresponding to the target vehicle.

[0095] In the embodiment of the present invention, optionally, the image depth information corresponding to the target vehicle may include depth information of the current environment corresponding to the target vehicle.

[0096] 103. Perform a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle.

[0097] In an embodiment of the present invention, optionally, the vehicle motion trajectory corresponding to the target vehicle may include a predicted motion trajectory of the target vehicle within a preset future time period; wherein the predicted motion trajectory includes a total displacement and a motion path of the target vehicle.

[0098] 104. Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle.

[0099] In an embodiment of the present invention, optionally, for example, visual information of the surrounding environment is collected by an on-board camera of the target vehicle, and features of objects, road markings, etc. in the environment are extracted using a visual odometry (VO) or visual SLAM technology, and sensor fusion is performed in combination with RTK information. Optimized position information is obtained through Kalman filtering and other technologies, and positioning compensation operations are performed on the target vehicle based on the optimized position information to obtain a high-precision positioning result of the target vehicle.

[0100] In an embodiment of the present invention, optionally, the current positioning information corresponding to the target vehicle can be acquired through one or more of an RTK receiver and a GNSS receiver; wherein, the RTK receiver provides high-precision position information, and the error is eliminated through differential correction; wherein, the RTK receiver is a high-precision global satellite positioning system (GNSS) technical equipment, which can provide centimeter-level positioning accuracy and is real-time; GNSS is the abbreviation of Global Navigation Satellite System, which includes global, regional and enhanced satellite navigation systems.

[0101] It can be seen that implementation Figure 1 The described positioning compensation method can collect vehicle image data corresponding to the target vehicle and extract image feature points, perform matching operations on all image feature points to obtain feature matching results, determine the image depth information corresponding to the target vehicle based on the feature matching results, perform trajectory prediction operations on the target vehicle to obtain the vehicle motion trajectory, and perform positioning compensation operations on the target vehicle based on the vehicle motion trajectory and the current positioning information of the target vehicle to obtain the corresponding positioning result. Through the fusion of vision and RTK, the accuracy and stability of the vehicle positioning system can be significantly improved. When the RTK signal is normal, centimeter-level high-precision positioning is provided, which is suitable for precise navigation. When the RTK signal is lost or unstable, the visual sensor can be combined with the IMU to maintain sub-meter-level positioning accuracy to ensure positioning The continuity and reliability of information, the fusion of multi-sensor information, and the optimization of data fusion algorithms such as Kalman filtering can effectively eliminate the noise and errors of each sensor, making the positioning results more stable, and can also obtain the characteristic information of the surrounding environment and maintain positioning. Even if the RTK signal fails, the system can still provide relative position information through vision. In an environment with poor RTK signals, the system can quickly switch to visual assisted positioning, and there will be no long-term positioning failure. The sensor fusion technology can provide position information to the visual system in a short time through the high-frequency attitude data provided by the IMU, ensuring the real-time and smoothness of positioning, and can also improve the positioning compensation accuracy of the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0102] Embodiment 2

[0103] See also Figure 2 , Figure 2 is a flow chart of another positioning compensation method disclosed in an embodiment of the present invention. Figure 2 The positioning compensation method described can be applied to a positioning compensation device or to the vehicle itself, wherein the positioning compensation device can be integrated in a cloud server or a local server, which is not limited in the embodiments of the present invention. Figure 2 As shown, the positioning compensation method may include the following operations:

[0104] 201. Collect vehicle image data corresponding to the target vehicle, extract image feature points in the vehicle image data, and perform feature point matching operations on all image feature points to obtain feature matching results.

[0105] 202. Based on the feature matching result, determine the image depth information corresponding to the target vehicle.

[0106] 203. Perform a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle.

[0107] 204. Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle.

[0108] In the embodiment of the present invention, for the detailed description of steps 201 to 204, please refer to the other descriptions of steps 101 to 104 in the first embodiment, and the embodiment of the present invention will not be repeated.

[0109] 205. Obtain motion data of the target vehicle.

[0110] In the embodiment of the present invention, the motion data of the target vehicle includes one or more of relative displacement data, relative direction data, motion speed data, and motion acceleration data of the target vehicle.

[0111] In the embodiment of the present invention, optionally, the motion data may also include a position vector, a velocity vector and posture information corresponding to the target vehicle.

[0112] 206. Generate relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory, and the current positioning information corresponding to the target vehicle.

[0113] In the embodiment of the present invention, optionally, the generating of the relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle may include:

[0114] Based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, the state vector of the target vehicle is determined, and the state of the target vehicle at the next moment is predicted through a preset state transition model, and the state of the target vehicle is predicted and corrected through a preset observation model to obtain a correction result, and the relative motion information of the target vehicle is generated based on the correction result;

[0115] The state vector of the target vehicle can be determined by the following formula:

[0116] x=[p vθ]^T

[0117] Where: p: position vector (x, y, z), v: velocity vector (vx, vy, vz), θ: attitude (roll, pitch, yaw).

[0118] 207. Perform a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information.

[0119] In the embodiment of the present invention, optionally, performing a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information may include:

[0120] According to the relative motion information, the target location corresponding to the target vehicle is determined, and the target location is updated to the location result corresponding to the target vehicle.

[0121] It can be seen that implementation Figure 2The described positioning compensation method can obtain the motion data of the target vehicle, generate the relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, and perform a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information. By combining the vehicle's motion data and the current positioning information, the real-time position of the vehicle can be more accurately predicted, thereby improving the accuracy and reliability of positioning. By obtaining the vehicle's relative displacement, direction, speed, acceleration and other motion data, the system can dynamically track the real-time status of the vehicle, which is particularly important for vehicle monitoring in a dynamic environment. In an environment where the RTK signal is poor, the system can quickly switch to visual-assisted positioning without long-term positioning failure. Sensor fusion technology can provide position information to the visual system in a short time through the high-frequency attitude data provided by the IMU, ensuring the real-time and smoothness of positioning. It can also help the system better understand the dynamic behavior of the vehicle through the generation of relative motion information, and help reduce error accumulation during long-term tracking through continuous positioning updates, and maintain the accuracy of positioning results. It not only improves the accuracy and real-time performance of positioning, but also enhances the dynamic tracking capability and safety of the system, and can further improve the positioning compensation accuracy of the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0122] In an optional embodiment, before performing a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle, the method further includes:

[0123] Acquire an image frame data set corresponding to the target vehicle, wherein the image frame data set includes a plurality of frames of image data corresponding to the target vehicle, and all the frames of image data are continuous image frames;

[0124] Performing a pose prediction operation on the image frame data set through a preset Kalman filter to obtain a pose prediction result corresponding to the image frame data set, and performing a data filtering operation on the pose prediction result to obtain a target pose result;

[0125] Among them, according to the image depth information, a trajectory prediction operation is performed on the target vehicle to obtain the vehicle motion trajectory corresponding to the target vehicle, including:

[0126] According to the image depth information and the target pose results, a trajectory prediction operation is performed on the target vehicle to obtain the vehicle motion trajectory corresponding to the target vehicle.

[0127] In this optional embodiment, optionally, a Kalman filter is an efficient recursive filter that can estimate the state of a dynamic system from a series of incomplete and noisy measurements. This filter plays an important role in real-time state estimation of dynamic systems with its high efficiency and recursive characteristics, and is widely used in control engineering, artificial intelligence, signal processing and other fields; the Kalman filter is based on two key steps: prediction and update. Through these two steps, it continuously recursively estimates the system state. In the prediction stage, the state of the next moment is predicted through the known system state transfer equation. In the update stage, the predicted state is corrected according to the new measurement data.

[0128] In this optional embodiment, optionally, the image frame data set includes a plurality of frames of image data corresponding to the target vehicle collected within a preset target time period; further, all the frames of image data are collected in a preset order.

[0129] In this optional embodiment, for example, a Kalman filter is used to predict and correct the pose. The Kalman filter uses the pose estimate of the current frame and the prediction result of the previous frame to smooth the motion path, reduce the influence of noise and error, and then obtain the target pose result.

[0130] In this optional embodiment, optionally, performing a trajectory prediction operation on the target vehicle according to the image depth information and the target pose result to obtain a vehicle motion trajectory corresponding to the target vehicle may include:

[0131] According to the image depth information and the target posture results, the relative posture changes of each frame image data corresponding to the target vehicle are determined, and based on all relative posture changes, the displacement posture changes and path posture changes of the target vehicle are calculated, and the vehicle motion trajectory corresponding to the target vehicle is determined based on the displacement posture changes and path posture changes.

[0132] It can be seen that the implementation of this optional embodiment can obtain the image frame data set corresponding to the target vehicle, perform a posture prediction operation on the image frame data set through a preset Kalman filter to obtain the corresponding posture prediction result, perform a data filtering operation on the posture prediction result to obtain the target posture result, and perform a trajectory prediction operation on the target vehicle according to the image depth information and the target posture result to obtain the vehicle motion trajectory corresponding to the target vehicle. Through continuous image frame data, the movement trend and change of the target vehicle can be captured more accurately, thereby improving the accuracy of trajectory prediction. The Kalman filter is used to perform posture prediction on the image frame data set, and the target can be tracked in real time. The dynamic changes of the vehicle enhance the dynamic tracking capability of the system. By performing data filtering operations on the pose prediction results, noise interference can be effectively reduced, and the stability and reliability of pose estimation can be improved. The pre-set Kalman filter can reduce the consumption of computing resources and improve the efficiency of data processing. By combining image depth information and pose prediction, not only the accuracy and robustness of trajectory prediction are improved, but also the accuracy and real-time performance of positioning are improved, and the dynamic tracking capability and safety of the system are enhanced. Furthermore, the positioning compensation accuracy of the target vehicle can be improved, and the accuracy of vehicle positioning can be improved, while the real-time navigation and positioning capabilities of the target vehicle can be enhanced.

[0133] In another optional embodiment, before determining the image depth information corresponding to the target vehicle based on the feature matching result, the method further includes:

[0134] Determine whether there are matching points to be eliminated in the feature matching results;

[0135] When it is determined that there are matching points to be eliminated in the feature matching results, based on a predetermined target algorithm, a data elimination operation is performed on the matching points to be eliminated in the feature matching results to update the feature matching results;

[0136] Analyze the updated feature matching results to obtain rotation matrix parameters and translation vector parameters corresponding to the feature matching results, and generate relative motion parameters based on the rotation matrix parameters and translation vector parameters;

[0137] Wherein, based on the feature matching result, the image depth information corresponding to the target vehicle is determined, including:

[0138] Based on the feature matching results and relative motion parameters, the image depth information corresponding to the target vehicle is determined.

[0139] In this optional embodiment, further optionally, when it is determined that there are no matching points to be eliminated in the feature matching results, the process can be terminated.

[0140] In this optional embodiment, optionally, the predetermined target algorithm may include a RANSAC algorithm; wherein, the RANSAC (RANdom SAmple Consensus) algorithm is an iterative method for estimating mathematical model parameters from a data set containing outliers; wherein, the core idea of ​​the RANSAC algorithm is to randomly select a subset of data to fit the model, and use this model to evaluate all data points to distinguish between inliers that conform to the model and outliers that do not conform to the model.

[0141] In this optional embodiment, optionally, there are no matching points to be eliminated in the updated feature matching results.

[0142] In this optional embodiment, optionally, analyzing the updated feature matching results to obtain rotation matrix parameters and translation vector parameters corresponding to the feature matching results, and generating relative motion parameters based on the rotation matrix parameters and translation vector parameters may include:

[0143] Analyze the updated feature matching results, solve the essential matrix by the five-point method, separate the rotation matrix parameters and the translation vector parameters corresponding to the feature matching results from the essential matrix, and generate relative motion parameters based on the rotation matrix parameters and the translation vector parameters; wherein the relative motion parameters include at least the rotation matrix parameters and the translation vector parameters;

[0144] Among them, the five-point method to solve the Essential Matrix is ​​a method used in computer vision to recover camera motion parameters from matching point pairs. The Essential Matrix contains the rotation and translation information between the two cameras.

[0145] It can be seen that the implementation of this optional embodiment can determine whether there are matching points to be eliminated in the feature matching results. If so, a data elimination operation is performed on the matching points to be eliminated in the feature matching results based on the target algorithm to update the feature matching results, and the updated feature matching results are analyzed to obtain corresponding rotation matrix parameters and translation vector parameters, and relative motion parameters are generated based on the rotation matrix parameters and the translation vector parameters. The image depth information corresponding to the target vehicle is determined based on the feature matching results and the relative motion parameters. By eliminating abnormal or erroneous matching points in the feature matching results, the matching accuracy can be significantly improved, which is crucial for the subsequent determination of image depth information and positioning and tracking of the target vehicle. By predetermining The target algorithm eliminates erroneous matching points, enhances the system's robustness to abnormal situations, and also enhances the stability of determining the image depth information corresponding to the target vehicle. The updated feature matching results can provide more accurate rotation matrix parameters and displacement vector parameters, which is beneficial to improve the accuracy and reliability of the subsequent generation of relative motion parameters and image depth information, and thus is beneficial to improve the accuracy of positioning compensation for the vehicle. It can also be dynamically adjusted according to different environments and conditions, enhancing the system's adaptability, so that it can maintain good performance in different application scenarios, and thus can also improve the accuracy of positioning compensation for the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0146] In yet another optional embodiment, based on the feature matching result and the relative motion parameter, determining the image depth information corresponding to the target vehicle, the determining the image depth information corresponding to the target vehicle includes:

[0147] Based on the feature matching result, at least one target frame data group is determined from all the frame image data, wherein the target frame data group includes at least two frame image data, and the corresponding order of all the frame image data included in each target frame data group is an adjacent order;

[0148] According to the relative motion parameters, feature point change information corresponding to each target frame data group is determined, and based on the feature point change information corresponding to all target frame data groups, a coordinate estimation operation is performed on the feature matching result to obtain spatial coordinate information corresponding to the feature matching result, and image depth information corresponding to the target vehicle is determined based on the spatial coordinate information;

[0149] The image depth information includes depth information corresponding to the current environment where the target vehicle is located.

[0150] In this optional embodiment, optionally, the feature point change information corresponding to each target frame data group may include one or more of coordinate change information and displacement change information of each target frame data group.

[0151] In this optional embodiment, optionally, performing a coordinate estimation operation on the feature matching result based on the feature point change information corresponding to all target frame data groups to obtain the spatial coordinate information corresponding to the feature matching result may include:

[0152] Based on the feature point change information corresponding to all target frame data groups, a coordinate estimation operation is performed on the feature matching result through a triangulation algorithm to obtain spatial coordinate information corresponding to the feature matching result, wherein the spatial coordinate information includes 3D coordinate values.

[0153] In this optional embodiment, optionally, triangulation estimation of 3D coordinates of feature points is a commonly used technique in computer vision, particularly in the fields of stereo vision and SLAM (Simultaneous Localization and Mapping); wherein the triangulation estimation includes the steps of feature point matching, camera pose estimation, constructing linear equations, and solving 3D coordinates.

[0154] In this optional embodiment, for example, the RANSAC algorithm is used to remove erroneous matching points, retain those feature point pairs that meet the model, reduce noise interference, and use the matched feature points to solve the essential matrix using the five-point method, and decompose the camera's rotation matrix R and translation vector t from the essential matrix, that is, relative motion, and triangulate and estimate the 3D coordinates of the feature points through the changes in the feature points in adjacent frames to obtain the depth information in the scene (even for a monocular camera, the depth can be restored through motion). Further optionally, the depth information in the scene refers to the distance information of each object in the scene relative to the observation point (usually the camera).

[0155] It can be seen that the implementation of this optional embodiment can determine at least one target frame data group in all frame image data based on the feature matching results, and determine the feature point change information corresponding to each target frame data group based on the relative motion parameters, thereby performing a coordinate estimation operation on the feature matching results to obtain the spatial coordinate information corresponding to the feature matching results, and determine the image depth information corresponding to the target vehicle based on the spatial coordinate information. By determining at least one target frame data group in all frame image data and determining the feature point change information based on the relative motion parameters, the 3D coordinates of the feature points can be more accurately estimated, thereby obtaining more accurate image depth information, and by analyzing the feature point changes of multiple adjacent frame image data, the error caused by matching a single frame or abnormal points is reduced. The depth information error caused by the error enhances the robustness of the system. By targeted selection of adjacent frame image data, unnecessary calculations can be reduced, the utilization of computing resources can be optimized, and the processing speed and efficiency can be improved. In addition, the consistency of the feature point change information obtained through adjacent frame image data can be improved, which is beneficial to improve the accuracy and reliability of the generated image depth information, thereby helping to improve the accuracy of positioning compensation for the vehicle. It can also be dynamically adjusted according to different environments and conditions, enhancing the system's adaptability so that it can maintain good performance in different application scenarios, thereby improving the positioning compensation accuracy of the target vehicle and improving the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0156] In yet another optional embodiment, based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle, including:

[0157] Based on the vehicle motion trajectory and the current positioning information of the target vehicle, the relative position change information of the target vehicle is determined;

[0158] According to the relative posture change information, the relative displacement change information of the target vehicle and the relative path change information of the target vehicle are determined, and according to the relative displacement change information and the relative path change information, the positioning result corresponding to the target vehicle is determined.

[0159] In this optional embodiment, optionally, the relative posture change information corresponding to the target vehicle may include one or more of relative displacement posture change information and relative rotation angle posture change information of the target vehicle.

[0160] In this optional embodiment, it is further optional to combine the IMU to provide high-frequency motion data to determine the positioning result corresponding to the target vehicle. For example, through the IMU motion data acquisition, the IMU provides the acceleration and angular velocity data of the vehicle at a higher sampling frequency. Even if the frame rate of the visual sensor is low, the IMU can still provide motion information in a short period of time, thereby filling the gap in the visual data; and through the fusion of IMU and visual data: through algorithms such as the extended Kalman filter (EKF), the high-frequency motion data of the IMU is fused with the relative displacement of the visual sensor. The IMU can reduce the jitter and drift of the visual system in a short period of time, and the visual data can correct the cumulative error that may exist in the IMU; thus, in the fusion process, the relative displacement information of the visual sensor is dominant, and the posture change is estimated through the continuous change of the image feature points. The IMU data provides accurate motion information in a short period of time, especially when turning quickly or vibrating, the IMU can quickly reflect the change. Through the Kalman filter, the complementary characteristics of the visual and IMU data are fused to generate a more stable and accurate posture estimate, thereby obtaining the positioning result corresponding to the target vehicle.

[0161] It can be seen that the implementation of this optional embodiment can determine the relative posture change information corresponding to the target vehicle based on the vehicle motion trajectory and the current positioning information of the target vehicle, determine the relative displacement change information and the relative path change information of the target vehicle according to the relative posture change information, and determine the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information. By determining the relative posture change information and the relative displacement change information of the target vehicle, the positioning error can be compensated more accurately, thereby improving the accuracy of the positioning result, and by analyzing multiple data sources (such as the vehicle motion trajectory and the current positioning information) to improve the robustness of the system, and to maintain stable positioning performance, by optimizing the algorithm and the compensation model, the accuracy and reliability of positioning can be improved without increasing additional hardware costs and energy consumption, and by comprehensively analyzing the vehicle motion trajectory and positioning information, not only the accuracy of the target vehicle positioning is improved, but also the robustness and adaptability of the system are enhanced, thereby also improving the accuracy of the subsequent positioning compensation of the target vehicle and improving the accuracy of the vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0162] In yet another optional embodiment, determining a positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information includes:

[0163] Obtain the environmental information of the current environment where the target vehicle is located, and build an environmental map corresponding to the current environment based on the environmental information;

[0164] Extract environmental feature information from the environmental map, and determine the position change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information;

[0165] A feature point matching operation is performed on the environmental feature information and the posture change feature information to obtain the posture feature matching result corresponding to the target vehicle, and the positioning result corresponding to the target vehicle is determined based on the posture feature matching result.

[0166] In this optional embodiment, optionally, the environmental information of the current environment where the target vehicle is located may be acquired in real time and may be acquired through a visual sensor.

[0167] In this optional embodiment, further optionally, after constructing an environmental map corresponding to the current environment based on environmental information, the above method may also include: identifying whether the current scene matches the known scene through loop detection technology, and if so, triggering the execution of extracting environmental feature information in the environmental map, and determining the posture change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information; if not, the process may be terminated.

[0168] In this optional embodiment, loop detection technology is a key concept in SLAM (Simultaneous Localization and Mapping). It refers to the process of identifying the robot's return to a previously visited location during the SLAM process; the main purpose of loop detection is to correct the accumulated navigation error. As the robot moves, its estimate of its own position tends to gradually deviate from the actual position. By identifying known locations, the system can adjust the map and the estimate of the robot's position, improving the accuracy of the entire system.

[0169] In this optional embodiment, for example, in a visual SLAM system, if an environmental map has been constructed before, the system can use loop detection technology to identify whether the current scene matches the previously known scene. When the visual system identifies that the vehicle enters an area that has been modeled, loop detection triggers repositioning to correct the accumulated errors caused by drift or sensor noise, and when the previous environmental map exists, the visual system will match the feature points in the current image with the known feature points in the map. Through these matching points, the system can calculate the absolute position information of the current vehicle. If the map accuracy is high, the repositioning process will quickly restore the precise position of the vehicle. Even in the case of poor or lost RTK signals, the vehicle can still be stably positioned based on visual information; further, when the vehicle re-enters the RTK signal coverage area, the system will obtain the absolute position information provided by RTK again. Data fusion and calibration: At this time, the visual system and RTK data are recalibrated through algorithms such as extended Kalman filters. The absolute positioning of RTK is used to correct the accumulated errors of the visual system, and the visual data continues to provide information on relative displacement. In this way, it can quickly recover to a high-precision absolute positioning state when the RTK signal is restored. When the vehicle re-enters the RTK signal coverage area, the system automatically restores the RTK positioning function and recalibrates the visual and RTK data through a fusion algorithm to ensure the consistency of the position information. The visual SLAM (Simultaneous Localization and Mapping) system is a technology used in the field of robotics or autonomous driving to simultaneously build environmental maps and position itself.

[0170] It can be seen that the implementation of this optional embodiment can obtain the environmental information of the current environment where the target vehicle is located and construct a corresponding environmental map, extract the environmental feature information in the environmental map, determine the corresponding posture change feature information of the target vehicle based on the relative displacement change information and the relative path change information, perform feature point matching operations on the environmental feature information and the posture change feature information to obtain the posture feature matching result corresponding to the target vehicle and then determine the corresponding positioning result of the target vehicle, and can use the static features in the environment (such as road markings, buildings, street lights, etc.) for auxiliary positioning, effectively solving the problem of unstable positioning of RTK in urban canyons, and RTK cannot receive satellite signals in tunnels and closed spaces, while the camera can maintain the displacement of the vehicle by tracking feature points in the tunnel (such as lights, road signs, etc.). Calculation ensures the continuity of positioning information, and can more accurately determine the position and posture changes of the target vehicle by constructing an environmental map and extracting environmental feature information, thereby improving the accuracy of the positioning results. By combining environmental feature information and position and posture change feature information, the system's robustness to abnormal situations is enhanced, so that the system can still maintain stable positioning performance when facing environmental changes or sensor noise. Accurate position and posture change feature information can improve the reliability of target vehicle tracking, which is beneficial to improving the accuracy and reliability of intelligent positioning of the target vehicle, and further helps to improve the accuracy and accuracy of positioning compensation for the target vehicle, thereby improving the accuracy of subsequent positioning compensation for the target vehicle and improving the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0171] Embodiment 3

[0172] See also Figure 3 , Figure 3 Schematic diagram of the structure of a positioning compensation device disclosed in an embodiment of the present invention. Figure 3 As shown, the positioning compensation device may include:

[0173] The acquisition module 301 is used to acquire vehicle image data corresponding to the target vehicle;

[0174] An extraction module 302 is used to extract image feature points from vehicle image data;

[0175] The matching module 303 is used to perform a feature point matching operation on all image feature points to obtain a feature matching result;

[0176] A determination module 304 is used to determine image depth information corresponding to the target vehicle based on the feature matching result;

[0177] The prediction module 305 is used to perform a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle;

[0178] The positioning module 306 is used to perform a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle.

[0179] It can be seen that implementation Figure 3 The described device can collect vehicle image data corresponding to the target vehicle and extract image feature points, perform matching operations on all image feature points to obtain feature matching results, determine the image depth information corresponding to the target vehicle based on the feature matching results, perform trajectory prediction operations on the target vehicle to obtain the vehicle motion trajectory, perform positioning compensation operations on the target vehicle based on the vehicle motion trajectory and the current positioning information of the target vehicle to obtain the corresponding positioning result. Through the fusion of vision and RTK, the accuracy and stability of the vehicle positioning system can be significantly improved. When the RTK signal is normal, centimeter-level high-precision positioning is provided, which is suitable for precise navigation. When the RTK signal is lost or unstable, the visual sensor can be combined with the IMU to maintain sub-meter-level positioning accuracy, ensuring positioning information The continuity and reliability of multi-sensor information can be integrated through Kalman filtering and other data fusion algorithm optimization to effectively eliminate the noise and error of each sensor, making the positioning result more stable. It can also obtain the characteristic information of the surrounding environment and maintain positioning. Even if the RTK signal fails, the system can still provide relative position information through vision. In an environment with poor RTK signal, the system can quickly switch to vision-assisted positioning, and there will be no long-term positioning failure. The sensor fusion technology can provide position information to the visual system in a short time through the high-frequency attitude data provided by IMU, ensuring the real-time and smoothness of positioning, and can also improve the positioning compensation accuracy of the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0180] In an optional embodiment, if Figure 4 As shown, the device also includes:

[0181] The acquisition module 307 is used to acquire the motion data of the target vehicle after the positioning module 306 performs a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, and obtains the positioning result corresponding to the target vehicle, wherein the motion data of the target vehicle includes one or more of the relative displacement data, relative direction data, motion speed data, and motion acceleration data of the target vehicle;

[0182] A generating module 308, for generating relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle;

[0183] The updating module 309 is used to perform a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information.

[0184] It can be seen that implementation Figure 4 The described device can obtain the motion data of the target vehicle, generate the relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, and perform a positioning update operation on the positioning result corresponding to the target vehicle according to the relative motion information. By combining the vehicle's motion data and the current positioning information, the real-time position of the vehicle can be more accurately predicted, thereby improving the accuracy and reliability of positioning. By obtaining the vehicle's relative displacement, direction, speed, acceleration and other motion data, the system can dynamically track the real-time status of the vehicle, which is particularly important for vehicle monitoring in a dynamic environment. In an environment with poor RTK signals, the system can quickly switch to visual-assisted positioning without long-term positioning failure. Sensor fusion technology can provide position information to the visual system in a short time through the high-frequency attitude data provided by the IMU, ensuring the real-time and smoothness of positioning. It can also help the system better understand the dynamic behavior of the vehicle through the generation of relative motion information, and help reduce error accumulation during long-term tracking through continuous positioning updates, and maintain the accuracy of positioning results. It not only improves the accuracy and real-time performance of positioning, but also enhances the dynamic tracking capability and safety of the system, and can further improve the positioning compensation accuracy of the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0185] In another optional embodiment, Figure 4 As shown, the acquisition module 307 is also used to acquire an image frame data set corresponding to the target vehicle before the prediction module 305 performs a trajectory prediction operation on the target vehicle according to the image depth information to obtain the vehicle motion trajectory corresponding to the target vehicle, wherein the image frame data set includes a plurality of frame image data corresponding to the target vehicle, and all frame image data are continuous image frames;

[0186] The prediction module 305 is further used to perform a posture prediction operation on the image frame data set through a preset Kalman filter to obtain a posture prediction result corresponding to the image frame data set, and perform a data filtering operation on the posture prediction result to obtain a target posture result;

[0187] The prediction module 305 performs a trajectory prediction operation on the target vehicle according to the image depth information, and obtains the vehicle motion trajectory corresponding to the target vehicle in the following specific manners:

[0188] According to the image depth information and the target pose results, a trajectory prediction operation is performed on the target vehicle to obtain the vehicle motion trajectory corresponding to the target vehicle.

[0189] It can be seen that implementation Figure 4The described device can obtain the image frame data set corresponding to the target vehicle, perform a posture prediction operation on the image frame data set through a preset Kalman filter to obtain the corresponding posture prediction result, perform a data filtering operation on the posture prediction result to obtain the target posture result, and perform a trajectory prediction operation on the target vehicle according to the image depth information and the target posture result to obtain the vehicle motion trajectory corresponding to the target vehicle. Through continuous image frame data, the movement trend and change of the target vehicle can be captured more accurately, thereby improving the accuracy of trajectory prediction. By using the Kalman filter to perform posture prediction on the image frame data set, the dynamic changes of the target vehicle can be tracked in real time, and the dynamic tracking capability of the system is enhanced. By performing a data filtering operation on the posture prediction result, noise interference can be effectively reduced, and the stability and reliability of posture estimation can be improved. In addition, through the preset Kalman filter, the consumption of computing resources can be reduced, and the efficiency of data processing can be improved. By combining image depth information and posture prediction, not only the accuracy and robustness of trajectory prediction are improved, but also the accuracy and real-time performance of positioning are improved, and the dynamic tracking capability and safety of the system are enhanced. Furthermore, the positioning compensation accuracy of the target vehicle and the accuracy of vehicle positioning can be improved, while the real-time navigation and positioning capability of the target vehicle can be enhanced.

[0190] In yet another optional embodiment, Figure 4 As shown, the device also includes:

[0191] The judging module 310 is used to judge whether there are matching points to be removed in the feature matching results before the determining module 304 determines the image depth information corresponding to the target vehicle based on the feature matching results;

[0192] The elimination module 311 is used for performing a data elimination operation on the matching points to be eliminated in the feature matching result based on a predetermined target algorithm when the judgment module 310 judges that there are matching points to be eliminated in the feature matching result, so as to update the feature matching result;

[0193] The generation module 308 is further used to analyze the updated feature matching results, obtain the rotation matrix parameters and the translation vector parameters corresponding to the feature matching results, and generate relative motion parameters based on the rotation matrix parameters and the translation vector parameters;

[0194] The specific manner in which the determination module 304 determines the image depth information corresponding to the target vehicle based on the feature matching result includes:

[0195] Based on the feature matching results and relative motion parameters, the image depth information corresponding to the target vehicle is determined.

[0196] It can be seen that implementation Figure 4The described device can determine whether there are matching points to be eliminated in the feature matching results. If so, a data elimination operation is performed on the matching points to be eliminated in the feature matching results based on a target algorithm to update the feature matching results. The updated feature matching results are analyzed to obtain corresponding rotation matrix parameters and translation vector parameters, and relative motion parameters are generated based on the rotation matrix parameters and the translation vector parameters. Image depth information corresponding to the target vehicle is determined based on the feature matching results and the relative motion parameters. The matching accuracy can be significantly improved by eliminating abnormal or erroneous matching points in the feature matching results, which is crucial for subsequent image depth information determination and target vehicle positioning and tracking. The algorithm eliminates erroneous matching points, enhances the system's robustness to abnormal situations, and also enhances the stability of determining the image depth information corresponding to the target vehicle. The updated feature matching results can provide more accurate rotation matrix parameters and displacement vector parameters, which is conducive to improving the accuracy and reliability of the subsequent generation of relative motion parameters and image depth information, and thus is also conducive to improving the accuracy of positioning compensation for the vehicle. It can also be dynamically adjusted according to different environments and conditions, enhancing the system's adaptability, so that it can maintain good performance in different application scenarios, and thus can also improve the accuracy of positioning compensation of the target vehicle and improve the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0197] In yet another optional embodiment, Figure 4 As shown, the determination module 304 determines the image depth information corresponding to the target vehicle based on the feature matching result and the relative motion parameter. The specific method of determining the image depth information corresponding to the target vehicle includes:

[0198] Based on the feature matching result, at least one target frame data group is determined from all the frame image data, wherein the target frame data group includes at least two frame image data, and the corresponding order of all the frame image data included in each target frame data group is an adjacent order;

[0199] According to the relative motion parameters, feature point change information corresponding to each target frame data group is determined, and based on the feature point change information corresponding to all target frame data groups, a coordinate estimation operation is performed on the feature matching result to obtain spatial coordinate information corresponding to the feature matching result, and image depth information corresponding to the target vehicle is determined based on the spatial coordinate information;

[0200] The image depth information includes depth information corresponding to the current environment where the target vehicle is located.

[0201] It can be seen that implementation Figure 4The described device can determine at least one target frame data group in all frame image data based on the feature matching results, and determine the feature point change information corresponding to each target frame data group based on the relative motion parameters, thereby performing a coordinate estimation operation on the feature matching results to obtain the spatial coordinate information corresponding to the feature matching results, and determine the image depth information corresponding to the target vehicle based on the spatial coordinate information. By determining at least one target frame data group in all frame image data and determining the feature point change information based on the relative motion parameters, the 3D coordinates of the feature points can be more accurately estimated, thereby obtaining more accurate image depth information, and by analyzing the feature point changes of multiple adjacent frame image data, the error caused by matching errors of a single frame or abnormal points is reduced. The depth information error caused by the adjacent frame image data is enhanced, and the robustness of the system is enhanced. By targetedly selecting adjacent frame image data, unnecessary calculations can be reduced, the utilization of computing resources can be optimized, and the processing speed and efficiency can be improved. In addition, the consistency of the feature point change information obtained through adjacent frame image data can be improved, which is beneficial to improving the accuracy and reliability of the generated image depth information, thereby helping to improve the accuracy of positioning compensation for the vehicle. It can also be dynamically adjusted according to different environments and conditions, and the system's adaptability can be enhanced, so that it can maintain good performance in different application scenarios, thereby improving the positioning compensation accuracy of the target vehicle and improving the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0202] In yet another optional embodiment, Figure 4 As shown, the positioning module 306 performs a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle. The specific method of obtaining the positioning result corresponding to the target vehicle includes:

[0203] Based on the vehicle motion trajectory and the current positioning information of the target vehicle, the relative position change information of the target vehicle is determined;

[0204] According to the relative posture change information, the relative displacement change information of the target vehicle and the relative path change information of the target vehicle are determined, and according to the relative displacement change information and the relative path change information, the positioning result corresponding to the target vehicle is determined.

[0205] It can be seen that implementation Figure 4The described device can determine the relative posture change information corresponding to the target vehicle based on the vehicle motion trajectory and the current positioning information of the target vehicle, determine the relative displacement change information and the relative path change information of the target vehicle according to the relative posture change information, and determine the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information. By determining the relative posture change information and the relative displacement change information of the target vehicle, the positioning error can be compensated more accurately, thereby improving the accuracy of the positioning result, and by analyzing multiple data sources (such as the vehicle motion trajectory and the current positioning information) to improve the robustness of the system, and to maintain stable positioning performance, by optimizing the algorithm and the compensation model, the accuracy and reliability of positioning can be improved without increasing additional hardware costs and energy consumption, by comprehensively analyzing the vehicle motion trajectory and positioning information, not only the accuracy of the target vehicle positioning is improved, but also the robustness and adaptability of the system are enhanced, thereby also improving the accuracy of the subsequent positioning compensation of the target vehicle and improving the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0206] In yet another optional embodiment, Figure 4 As shown, the specific manner in which the positioning module 306 determines the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information includes:

[0207] Obtain the environmental information of the current environment where the target vehicle is located, and build an environmental map corresponding to the current environment based on the environmental information;

[0208] Extract environmental feature information from the environmental map, and determine the position change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information;

[0209] A feature point matching operation is performed on the environmental feature information and the posture change feature information to obtain the posture feature matching result corresponding to the target vehicle, and the positioning result corresponding to the target vehicle is determined based on the posture feature matching result.

[0210] It can be seen that implementation Figure 4The described device can obtain environmental information of the current environment where the target vehicle is located and construct a corresponding environmental map, extract environmental feature information in the environmental map, determine the corresponding posture change feature information of the target vehicle based on the relative displacement change information and the relative path change information, perform feature point matching operations on the environmental feature information and the posture change feature information to obtain the posture feature matching result corresponding to the target vehicle and then determine the corresponding positioning result of the target vehicle, and can use static features in the environment (such as road markings, buildings, street lights, etc.) for auxiliary positioning, effectively solving the problem of unstable positioning of RTK in urban canyons, and RTK cannot receive satellite signals in tunnels and closed spaces, while the camera can maintain the displacement calculation of the vehicle by tracking feature points in the tunnel (such as lights, road signs, etc.). In addition to ensuring the continuity of positioning information, the position and posture changes of the target vehicle can be determined more accurately by constructing an environmental map and extracting environmental feature information, thereby improving the accuracy of the positioning results. By combining environmental feature information and position and posture change feature information, the system's robustness to abnormal situations is enhanced, so that the system can still maintain stable positioning performance when facing environmental changes or sensor noise. Accurate position and posture change feature information can improve the reliability of target vehicle tracking, which is beneficial to improving the accuracy and reliability of intelligent positioning of the target vehicle, and further helps to improve the accuracy and accuracy of positioning compensation for the target vehicle, thereby improving the accuracy of subsequent positioning compensation for the target vehicle and improving the accuracy of vehicle positioning, while enhancing the real-time navigation and positioning capabilities of the target vehicle.

[0211] Embodiment 4

[0212] See also Figure 5 , Figure 5 FIG. 1 is a schematic diagram of the structure of another positioning compensation device disclosed in an embodiment of the present invention. Figure 5 As shown, the positioning compensation device may include:

[0213] A memory 401 storing executable program codes;

[0214] a processor 402 coupled to the memory 401;

[0215] The processor 402 calls the executable program code stored in the memory 401 to execute the steps of the positioning compensation method described in the first embodiment of the present invention or the second embodiment of the present invention.

[0216] Embodiment 5

[0217] An embodiment of the present invention discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute the steps in the positioning compensation method described in Embodiment 1 or Embodiment 2 of the present invention.

[0218] Embodiment 6

[0219] An embodiment of the present invention discloses a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to enable a computer to execute the steps in the positioning compensation method described in Embodiment 1 or Embodiment 2.

[0220] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0221] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, a magnetic disk storage, a magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0222] Finally, it should be noted that the positioning compensation method and device disclosed in the embodiments of the present invention only disclose the preferred embodiments of the present invention, which are only used to illustrate the technical solution of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A positioning compensation method, characterized in that: The method comprises: Collecting vehicle image data corresponding to the target vehicle, extracting image feature points in the vehicle image data, and performing a feature point matching operation on all the image feature points to obtain a feature matching result; Based on the feature matching result, determining the image depth information corresponding to the target vehicle; According to the image depth information, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle; Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, a positioning compensation operation is performed on the target vehicle to obtain a positioning result corresponding to the target vehicle.

2. The positioning compensation method according to claim 1, characterized in that: After performing a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle, the method further includes: Acquiring motion data of the target vehicle, wherein the motion data of the target vehicle includes one or more of relative displacement data, relative direction data, motion speed data, and motion acceleration data of the target vehicle; Generate relative motion information of the target vehicle based on the motion data, the vehicle motion trajectory and the current positioning information corresponding to the target vehicle; According to the relative motion information, a positioning update operation is performed on the positioning result corresponding to the target vehicle.

3. The positioning compensation method according to claim 2, characterized in that: Before performing a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle, the method further includes: Acquire an image frame data set corresponding to the target vehicle, wherein the image frame data set includes a plurality of frames of image data corresponding to the target vehicle, and all of the frame image data are continuous image frames; Performing a posture prediction operation on the image frame data set through a preset Kalman filter to obtain a posture prediction result corresponding to the image frame data set, and performing a data filtering operation on the posture prediction result to obtain a target posture result; The step of performing a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle includes: According to the image depth information and the target pose result, a trajectory prediction operation is performed on the target vehicle to obtain a vehicle motion trajectory corresponding to the target vehicle.

4. The positioning compensation method according to claim 2, characterized in that: Before determining the image depth information corresponding to the target vehicle based on the feature matching result, the method further includes: Determine whether there are matching points to be eliminated in the feature matching results; When it is determined that the matching point to be eliminated exists in the feature matching result, based on a predetermined target algorithm, a data elimination operation is performed on the matching point to be eliminated in the feature matching result to update the feature matching result; Analyze the updated feature matching result to obtain rotation matrix parameters and translation vector parameters corresponding to the feature matching result, and generate relative motion parameters based on the rotation matrix parameters and the translation vector parameters; Wherein, determining the image depth information corresponding to the target vehicle based on the feature matching result includes: Based on the feature matching result and the relative motion parameter, image depth information corresponding to the target vehicle is determined.

5. The positioning compensation method according to claim 4, characterized in that: The step of determining the image depth information corresponding to the target vehicle based on the feature matching result and the relative motion parameter comprises: Based on the feature matching result, at least one target frame data group is determined from all the frame image data, wherein the target frame data group includes at least two frame image data, and the corresponding order of all the frame image data included in each target frame data group is an adjacent order; Determine feature point change information corresponding to each target frame data group according to the relative motion parameters, and perform a coordinate estimation operation on the feature matching result based on the feature point change information corresponding to all the target frame data groups to obtain spatial coordinate information corresponding to the feature matching result, and determine image depth information corresponding to the target vehicle based on the spatial coordinate information; The image depth information includes depth information corresponding to the current environment where the target vehicle is located.

6. The positioning compensation method according to claim 1, characterized in that: The performing a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle includes: Based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle, determining the relative position change information corresponding to the target vehicle; According to the relative posture change information, the relative displacement change information of the target vehicle and the relative path change information of the target vehicle are determined, and according to the relative displacement change information and the relative path change information, the positioning result corresponding to the target vehicle is determined.

7. The positioning compensation method according to claim 6, characterized in that: Determining the positioning result corresponding to the target vehicle according to the relative displacement change information and the relative path change information includes: Acquire environmental information of the current environment where the target vehicle is located, and construct an environmental map corresponding to the current environment based on the environmental information; Extracting environmental feature information from the environmental map, and determining position change feature information corresponding to the target vehicle based on the relative displacement change information and the relative path change information; A feature point matching operation is performed on the environmental feature information and the posture change feature information to obtain a posture feature matching result corresponding to the target vehicle, and a positioning result corresponding to the target vehicle is determined based on the posture feature matching result.

8. A positioning compensation device, characterized in that: The device comprises: An acquisition module, used for acquiring vehicle image data corresponding to a target vehicle; An extraction module, used to extract image feature points from the vehicle image data; A matching module, used to perform a feature point matching operation on all the image feature points to obtain a feature matching result; A determination module, used to determine the image depth information corresponding to the target vehicle based on the feature matching result; A prediction module, configured to perform a trajectory prediction operation on the target vehicle according to the image depth information to obtain a vehicle motion trajectory corresponding to the target vehicle; The positioning module is used to perform a positioning compensation operation on the target vehicle based on the vehicle motion trajectory and the current positioning information corresponding to the target vehicle to obtain a positioning result corresponding to the target vehicle.

9. A positioning compensation device, characterized in that: The device comprises: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the positioning compensation method as described in any one of claims 1-7.

10. A computer storage medium, characterized in that: The computer storage medium stores computer instructions, and when the computer instructions are called, they are used to execute the positioning compensation method as described in any one of claims 1-7.

Images