Vehicle positioning method and device for multi-layer space transition area, equipment and medium

Abstract

The invention relates to a vehicle positioning method, device and equipment for a multi-layer space transition area and a medium, and belongs to the technical field of vehicle localization, the method comprises the following steps: dispersing the multi-layer space transition area into sequence nodes, and collecting vehicle pose information, aerial view image information and altitude information corresponding to each node; collecting vehicle position information, associating the vehicle position information with the vehicle pose information, the aerial view image information and the altitude information, and constructing a pose feature map oriented to the multi-layer space transition area; a DTW algorithm is utilized to be matched with the pose feature map, and a pose feature candidate set is determined; the nodes in the pose feature candidate set are used as particles, the real-time vehicle pose features are used as observation information, the likelihood probability between the observation information and the particles is calculated, and the optimal matching nodes in the pose feature candidate set are output; and calculating the pose of the vehicle in the global coordinate system based on the optimal matching node. The method is used for improving high-precision and robust positioning of a vehicle in a transition area.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle positioning, and in particular to a vehicle positioning method, device, equipment and medium in a multi-layer space transition area. Background Art

[0002] Multi-storey parking lots are one of the important scenes for urban parking. The transition area is the "throat" connecting indoor and outdoor areas, which directly determines whether the last "kilometer" of the vehicle can achieve intelligent driving. The transition area is often a structural degradation scene with special properties such as sudden changes in light, poor texture and structural features, turns and ramps, and is a satellite signal blind area, making it difficult to obtain effective vehicle location information.

[0003] Among the current positioning methods, V2X (Vehicle-to-Everything) has the disadvantages of high deployment and maintenance costs, SLAM (Simultaneous Localization and Mapping) and inertial navigation systems have the disadvantages of cumulative errors, and visual maps and laser maps have the characteristics of poor positioning accuracy and stability in multi-layer transition areas. These methods cannot effectively solve the positioning problem in transition areas.

[0004] Therefore, how to solve the positioning needs of vehicles in the "last mile" of the transition area is a technical problem that technicians in this field urgently need to solve. Summary of the invention

[0005] In view of this, it is necessary to provide a vehicle positioning method, device, equipment and medium in a multi-layer space transition area to improve the high-precision and robust positioning of the vehicle in the multi-layer space transition area.

[0006] In order to solve the above problems, in a first aspect, the present invention provides a vehicle positioning method in a multi-layer space transition area, comprising: Discretize the multi-layer spatial transition area into sequence nodes, and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; Collecting vehicle position information, associating the vehicle position information with the vehicle posture information, bird's-eye view image information and altitude information, and constructing a posture feature map database for the multi-layer spatial transition area; Use the DTW algorithm to match the pose feature map to determine the candidate set of pose features; Taking the nodes in the candidate set of posture features as particles and the real-time vehicle posture features as observation information, calculating the likelihood probability between the observation information and the particles, and outputting the best matching node in the candidate set of posture features; The position and posture of the vehicle in the global coordinate system is calculated based on the best matching node.

[0007] In a possible implementation, collecting vehicle posture information, bird's-eye view image information, and altitude information corresponding to each node includes: The barometer is used to collect the altitude information corresponding to each node, the IMU is used to collect the vehicle posture information, and the BEV is used to collect the bird's-eye view image information; The collecting of vehicle location information includes: Use INS to collect vehicle location information.

[0008] In a possible implementation, after collecting the bird's-eye view image information, the method further includes: Extracting lane line semantic information from the bird's-eye view image, and obtaining a deviation of the vehicle relative to the lane line based on the lane line semantic information; The vehicle posture information is normalized using the deviation of the vehicle relative to the lane line.

[0009] In a possible implementation, after collecting the vehicle location information, the method further includes: The structural information constraints of the multi-layer spatial transition region are obtained, and the vehicle position information is optimized based on the structural information constraints of the multi-layer spatial transition region.

[0010] In a possible implementation, associating the vehicle position information with the vehicle posture information, the bird's-eye view image information, and the altitude information to construct a posture feature map for the multi-layer spatial transition area includes: Traverse the multi-layer spatial transition area and convert the collected data into Organizing and storing, completing the construction of the posture feature map for the multi-layer space transition area, wherein: x ( i ), y ( i ) is the vehicle in i The vehicle location information corresponding to each node, a ( i ), r ( i ), p ( i ) is the vehicle in i The pose information at each node, h ( i ) is the vehicle in i The altitude information at each node, I ( i ) is the first i The bird's-eye view image information of each node, is the number of nodes.

[0011] In a possible implementation, before using the DTW algorithm to match the pose feature map to determine the pose feature candidate set, the method further includes: Construct a short-term uniform motion model based on the vehicle's historical position information; Determine a preliminary matching range of the vehicle in the posture feature map based on the short-term uniform motion model; The DTW algorithm is used to match the pose feature map to determine the pose feature candidate set, specifically including: The DTW algorithm is used to match the preliminary matching range in the pose feature map to determine the pose feature candidate set.

[0012] In a possible implementation, taking the nodes in the pose feature candidate set as particles, taking the real-time vehicle pose feature as observation information, calculating the likelihood probability between the observation information and the particles, and outputting the best matching node in the pose feature candidate set, includes: The likelihood probability between the real-time vehicle posture feature and each node in the posture feature candidate set is calculated, and the particles are sampled by importance to output the best matching node in the posture feature candidate set.

[0013] In a possible implementation, calculating the position and posture of the vehicle in the global coordinate system based on the best matching node includes: The relative position relationship between the real-time vehicle posture feature and the posture feature of each node in the best matching node is calculated, and the posture of the vehicle in the global coordinate system is solved through a preset coordinate transformation relationship.

[0014] In a second aspect, the present invention provides a vehicle positioning device in a multi-layer space transition area, comprising: An information collection module is used to discretize the multi-layer spatial transition area into sequence nodes and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; A map construction module, used for collecting vehicle position information, associating the vehicle position information with the vehicle posture information, bird's-eye view image information and altitude information, and constructing a posture feature map database for the multi-layer space transition area; A matching module is used to match the pose feature map using the DTW algorithm to determine a candidate set of pose features; A calculation module, used to use the nodes in the candidate set of posture features as particles and the real-time vehicle posture features as observation information, calculate the likelihood probability between the observation information and the particles, and output the best matching node in the candidate set of posture features; The vehicle positioning module is used to calculate the position and posture of the vehicle in the global coordinate system based on the best matching node.

[0015] In a third aspect, the present invention further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned method for positioning a vehicle in a multi-layer spatial transition area when executing the computer program.

[0016] In a fourth aspect, the present invention further provides a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the vehicle positioning method in the multi-layer spatial transition area are implemented as described above.

[0017] The beneficial effects of the present invention are: The present invention first discretizes the multi-layer space transition area into sequence nodes, and then uses various sensors to collect vehicle posture information, bird's-eye view image information and altitude information corresponding to each node and collects vehicle position information, and associates various information to construct a posture feature map for the multi-layer space transition area; by using sensors to collect information, the hardware deployment and maintenance costs are reduced, and the scene adaptability is improved; by associating the various collected information, robust positioning in multi-layer space structural degradation scenes can be achieved, and the positioning accuracy of the posture feature map is improved; the nodes on the posture feature map are used as particles, and the real-time vehicle posture features are used as observation information, and the DTW algorithm is used to determine the vehicle positioning result. By gradually narrowing the matching range of the vehicle on the posture feature map, the positioning accuracy of the vehicle positioning result is further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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 those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A method flow chart of an embodiment of a method for locating a vehicle in a multi-layer space transition area provided by the present invention; Figure 2 A schematic diagram of normalization of vehicle posture information provided by an embodiment of the present invention; Figure 3 A schematic diagram of constructing a posture feature map database provided by one embodiment of the present invention; Figure 4 A schematic diagram of vehicle posture feature sequence matching extracted by an embodiment of the present invention; Figure 5 A schematic structural diagram of an embodiment of a vehicle positioning device for a multi-layer space transition area provided by the present invention; Figure 6A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0021] The descriptions of "first", "second", etc. involved in the embodiments of the present invention are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the technical features defined as "first" and "second" may explicitly or implicitly include at least one of the features.

[0022] Reference to an "embodiment" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiment 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.

[0023] A specific embodiment of the present invention, as Figure 1 As shown, Figure 1 A method flow chart of an embodiment of a method for locating a vehicle in a multi-layer space transition area provided by the present invention includes: S101: discretize the multi-layer spatial transition area into sequence nodes, and collect vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; S102: Collecting vehicle position information, associating the vehicle position information with vehicle posture information, bird's-eye view image information and altitude information, and constructing a posture feature map database for multi-layer spatial transition areas; S103: Using the nodes on the posture feature map as particles and the real-time vehicle posture features as observation information, the DTW algorithm is used to determine the vehicle positioning result.

[0024] The present invention first discretizes the multi-layer space transition area into sequence nodes, and then uses various sensors to collect vehicle posture information, bird's-eye view image information and altitude information corresponding to each node and collects vehicle position information, and associates various information to construct a posture feature map for the multi-layer space transition area; by using sensors to collect information, the hardware deployment and maintenance costs are reduced, and the scene adaptability is improved; by associating the various collected information, robust positioning in multi-layer space structural degradation scenes can be achieved, and the positioning accuracy of the posture feature map is improved; the nodes on the posture feature map are used as particles, and the real-time vehicle posture features are used as observation information, and the DTW algorithm is used to determine the vehicle positioning result. By gradually narrowing the matching range of the vehicle on the posture feature map, the positioning accuracy of the vehicle positioning result is further improved.

[0025] In one embodiment of the present invention, collecting vehicle posture information, bird's-eye view image information and altitude information corresponding to each node includes: The barometer is used to collect the altitude information corresponding to each node, the IMU is used to collect the vehicle posture information, and the BEV is used to collect the bird's-eye view image information; Collect vehicle location information, including: Use INS to collect vehicle location information.

[0026] First of all, it should be noted that before collecting various information, the multi-source sensors IMU (Inertial Measurement Unit), INS (Integrated Navigation System) and BEV (Bird's Eye View) need to be calibrated for spatiotemporal consistency. Specifically, the trajectory consistency method is used to align the IMU and camera and the INS external parameter calibration to achieve spatial unification of different sensor data.

[0027] Among them, the sensor intrinsic calibration consists of two parts: IMU intrinsic parameters and camera intrinsic parameters. The IMU intrinsic parameters include axis deviation, scale factor error, constant zero bias, etc.; the camera intrinsic parameters include camera focal length, distortion coefficient, principal point coordinates and other parameters.

[0028] Sensor extrinsic calibration includes spatial consistency calibration and temporal consistency calibration. For spatial consistency, the extrinsic parameter calibration between IMU and camera is solved by the famous Kaliber, and the extrinsic parameter calibration between camera and INS is calibrated by trajectory consistency method, thereby achieving spatial consistency between heterogeneous sensors; for temporal consistency, linear interpolation method is used to achieve it.

[0029] In one embodiment of the present invention, after collecting the bird's-eye view image information, the method further includes: Extracting the lane line semantic information of the BEV image, and obtaining the deviation of the vehicle relative to the lane line based on the lane line semantic information; The vehicle's position information is normalized using the deviation of the vehicle relative to the lane line.

[0030] It should be noted that when using vehicle posture information to construct a vehicle posture feature map database, the vehicle posture information needs to be normalized. Specifically, BEVTransformer is used to extract the lane line semantic information of the bird's-eye view image, and based on the lane line semantic information, a linear regression method is used to obtain the vehicle's deviation relative to the lane line; then the deviation is used to correct the vehicle heading obtained by the IMU to achieve normalization of the vehicle posture information.

[0031] For example, Figure 2 As shown, Figure 2 A schematic diagram of normalizing vehicle posture information provided by an embodiment of the present invention. The lane line information in the bird's-eye view image is extracted, and the relative posture relationship with the lane line is calculated. Assume that "1" is the vehicle heading information stored in the posture map, and "2" is the actual heading of the vehicle during the positioning process. By calculating the angle between the vehicle and the lane line, such as θ , to achieve the normalization of the vehicle heading.

[0032] In one embodiment of the present invention, after collecting the vehicle location information, the method further includes: The structural information constraints of the multi-layer spatial transition region are obtained, and the vehicle position information is optimized based on the structural information constraints of the multi-layer spatial transition region.

[0033] It should be noted that the vehicle position information in the present invention is specifically high-precision vehicle position information. When using the vehicle position information to construct a vehicle posture feature map database, it is necessary to use the spatial structure information of the multi-layer transition area to optimize the position information collected by INS to obtain high-precision vehicle position information.

[0034] Specifically, by adding structural information constraints in the multi-layer space transition area to weaken the influence of cumulative errors, the position accuracy of the vehicle is further optimized. The structural information constraints in the multi-layer space transition area are the structural features of the multi-layer space transition area, such as the ramp structure at the entrance and exit, and the turning structure in the multi-layer space, as constraint information. By setting the vehicle posture as a node and the node constraint as an edge for optimization, the vehicle position accuracy is improved.

[0035] In one embodiment of the present invention, the vehicle position information is associated with the vehicle posture information, the bird's-eye view image information and the altitude information to construct a posture feature map for the multi-layer spatial transition area, including: Traverse the multi-layer spatial transition area and convert the collected data into Organize and store to complete the construction of the pose feature map for the multi-layer spatial transition area, where: x ( i ), y ( i ) is the vehicle in i The vehicle location information corresponding to each node, a ( i ), r ( i ), p ( i ) is the vehicle in i The pose information at each node, h ( i ) is the vehicle in i The altitude information at each node, I ( i ) is the first i The bird's-eye view image information of each node, is the number of nodes.

[0036] It can be understood that by obtaining the node images collected by the vehicle at each node position, the global descriptor, local descriptor and local feature points of each node image are calculated as the bird's-eye view image information, and associated with the vehicle position information, vehicle posture information and vehicle altitude information to construct a posture feature map database.

[0037] The data included in each node is: (1) x ( i ), y ( i ) is the vehicle in i The vehicle position corresponding to each node is a ( i ), r ( i ), p ( i ) is the vehicle in i The pose information at each node, h ( i ) is the vehicle in i The altitude information at each node, I ( i ) is the first i The bird's-eye view image information of each node. Traverse all the areas to be located in the multi-layer space, and organize and store the collected data according to the above formula to complete the construction of the pose feature map of the area to be located. Figure 3 As shown, Figure 3A schematic diagram of constructing a posture feature map database provided by one embodiment of the present invention.

[0038] In one embodiment of the present invention, before using the DTW algorithm to match the pose feature map to determine the pose feature candidate set, the method further includes: Construct a short-term uniform motion model based on the vehicle's historical position information; Determine the initial matching range of the vehicle in the posture feature map based on the short-term uniform motion model; It is understandable that, considering the complexity of the multi-layer structural transition area, the vehicle's historical position information can be used to predict the current vehicle position information to narrow the scope of the pose feature map matching and improve the map matching efficiency and stability. In the positioning process, the real-time vehicle pose features also need to be normalized. The normalization method is the same as the normalization method in constructing the pose feature map, which will not be repeated here.

[0039] Use the DTW algorithm to match the pose feature map to determine the pose feature candidate set, including: The DTW algorithm is used to match the preliminary matching range in the pose feature map to determine the candidate set of pose features.

[0040] The DTW algorithm calculates the distance information between the real-time acquisition sequence and the map node sequence, and then constructs a distance information matrix. It uses the distance information matrix to find the map node sequence that matches the real-time acquisition sequence, and determines the candidate set by setting a threshold.

[0041] It can be understood that within the preliminary matching range, the DTW algorithm is used to determine the candidate set for further matching of the map, that is, to match the latest acquired posture features with the historical posture features and the posture sequence in the posture feature map to determine the candidate set of posture features. This range is the coarse matching range, and its purpose is to improve the robustness and efficiency of map matching.

[0042] The map nodes in the pose feature map database are regarded as particles in the particle filter algorithm. The pose information and altitude information obtained by the vehicle are used as observations. The pose features are normalized by the lane line information detected by the BEV image. The normalized sequence pose and altitude information are used to match the best node sequence in the pose feature map database through the DTW algorithm. The similarity between nodes is measured using the cosine algorithm.

[0043] Specifically, for example, Figure 4 As shown, Figure 4A schematic diagram of a vehicle posture feature sequence matching extracted in one embodiment of the present invention. The triangles are historical vehicle posture features stored in the map, and the quadrilaterals are real-time vehicle posture features during positioning. Through the sliding window model, the historically acquired vehicle posture feature sequence is matched with the posture sequence stored in the map to determine the best map matching range, i.e., the posture feature candidate set.

[0044] In one embodiment of the present invention, the nodes in the pose feature candidate set are taken as particles, the real-time vehicle pose features are taken as observation information, the likelihood probability between the observation information and the particles is calculated, and the best matching node in the pose feature candidate set is output, including: The likelihood probability between the real-time vehicle posture feature and each node in the posture feature candidate set is calculated, and the particles are sampled by importance to output the best matching node in the posture feature candidate set.

[0045] It can be understood that based on the candidate set of posture features, the normalized posture and altitude information is further used to calculate the likelihood probability of the current vehicle posture information and each particle, and then the particles are sampled for importance. Based on the particle importance sampling, a part of the particles with the largest likelihood probability is selected as the best matching node.

[0046] In one embodiment of the present invention, solving the position and posture of the vehicle in the global coordinate system based on the best matching node includes: The relative position relationship between the real-time vehicle posture features and the posture features of each node in the best matching node is calculated, and the vehicle's posture in the global coordinate system is solved through the preset coordinate transformation relationship.

[0047] It can be understood that based on the best matching node on the map, the posture features acquired in real time are aligned with the posture features in the best matching node, and the relative position relationship between the posture features acquired in real time and the posture features in the best matching node is calculated, and then the posture of the vehicle in the global coordinate system is solved through coordinate transformation.

[0048] In order to better implement the vehicle positioning method in the multi-layer space transition area in the embodiment of the present invention, based on the vehicle positioning method in the multi-layer space transition area, correspondingly, Figure 5 As shown, the embodiment of the present invention further provides a vehicle positioning device in a multi-layer space transition area, and the vehicle positioning device 500 in a multi-layer space transition area includes: The information collection module 501 is used to discretize the multi-layer spatial transition area into sequence nodes, and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; A map construction module 502 is used to collect vehicle position information, associate the vehicle position information with vehicle posture information, bird's-eye view image information and altitude information, and construct a posture feature map database for multi-layer spatial transition areas; A matching module 503 is used to match the pose feature map using a DTW algorithm to determine a candidate pose feature set; The calculation module 504 is used to use the nodes in the pose feature candidate set as particles and the real-time vehicle pose feature as observation information, calculate the likelihood probability between the observation information and the particles, and output the best matching node in the pose feature candidate set; The vehicle positioning module 505 is used to calculate the position and posture of the vehicle in the global coordinate system based on the best matching node.

[0049] The vehicle positioning device 500 for the multi-layer space transition area provided in the above embodiment can implement the technical solution described in the above embodiment of the vehicle positioning method for the multi-layer space transition area. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above embodiment of the vehicle positioning method for the multi-layer space transition area, which will not be repeated here.

[0050] like Figure 6 As shown, the present invention also provides an electronic device 600. The electronic device 600 includes a processor 601, a memory 602 and a display 603. Figure 6 Only some components of the electronic device 600 are shown, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0051] In some embodiments, the processor 601 may be a central processing unit (CPU), a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 602, such as the vehicle positioning method in the multi-layer spatial transition area of ​​the present invention.

[0052] In some embodiments, the processor 601 may be a single server or a server group. The server group may be centralized or distributed. In some embodiments, the processor 601 may be local or remote. In some embodiments, the processor 601 may be implemented in a cloud platform. In one embodiment, the cloud platform may include a private cloud, a public cloud, a hybrid cloud, a community cloud, a distributed cloud, an internal cloud, a multi-cloud, etc., or any combination thereof.

[0053] In some embodiments, the memory 602 may be an internal storage unit of the electronic device 600, such as a hard disk or memory of the electronic device 600. In other embodiments, the memory 602 may also be an external storage device of the electronic device 600, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc., equipped on the electronic device 600.

[0054] Furthermore, the memory 602 may include both an internal storage unit of the electronic device 600 and an external storage device. The memory 602 is used to store application software installed in the electronic device 600 and various data.

[0055] In some embodiments, the display 603 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 603 is used to display information of the electronic device 600 and to display a visual user interface. The components 601-603 of the electronic device 600 communicate with each other through a system bus.

[0056] In some embodiments, when the processor 601 executes the vehicle positioning program for the multi-layer spatial transition area in the memory 602, the following steps may be implemented: Discretize the multi-layer spatial transition area into sequence nodes, and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; Collect vehicle location information, associate the vehicle location information with vehicle posture information, bird's-eye view image information and altitude information, and build a posture feature map database for multi-layer spatial transition areas; Use the DTW algorithm to match the pose feature map to determine the candidate set of pose features; Taking the nodes in the candidate set of pose features as particles and the real-time vehicle pose features as observation information, the likelihood probability between the observation information and the particles is calculated, and the best matching node in the candidate set of pose features is output; The vehicle's position and posture in the global coordinate system are solved based on the best matching node.

[0057] It should be understood that: when the processor 601 executes the vehicle positioning program in the multi-layer spatial transition area in the memory 602, in addition to the above functions, other functions can also be implemented. For details, please refer to the description of the corresponding method embodiment above.

[0058] Furthermore, the embodiment of the present invention does not specifically limit the type of the electronic device 600 mentioned, and the electronic device 600 may be a portable electronic device such as a mobile phone, a tablet computer, a personal digital assistant (PDA), a wearable device, a laptop computer, etc. Exemplary embodiments of portable electronic devices include but are not limited to portable electronic devices equipped with IOS, Android, Microsoft or other operating systems. The above-mentioned portable electronic devices may also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 600 may not be a portable electronic device, but a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0059] Accordingly, an embodiment of the present application also provides a computer-readable storage medium, which is used to store computer-readable programs or instructions. When the program or instructions are executed by a processor, it can implement the steps or functions of the vehicle positioning method in the multi-layer spatial transition area provided in the above-mentioned method embodiments.

[0060] Those skilled in the art can understand that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0061] The above are only preferred specific implementations of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A vehicle positioning method in a multi-layer space transition area, characterized in that: include: Discretize the multi-layer spatial transition area into sequence nodes, and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; Collecting vehicle position information, associating the vehicle position information with the vehicle posture information, the bird's-eye view image information and the altitude information, and constructing a posture feature map for the multi-layer space transition area; Use the DTW algorithm to match the pose feature map to determine the candidate set of pose features; Taking nodes in the candidate set of posture features as particles and real-time vehicle posture features as observation information, calculating the likelihood probability between the observation information and the particles, and outputting the best matching node in the candidate set of posture features; The position and posture of the vehicle in the global coordinate system is calculated based on the best matching node.

2. The vehicle positioning method in a multi-layer space transition area according to claim 1, characterized in that: After collecting the bird's-eye view image information, the method further includes: Extracting lane line semantic information from the bird's-eye view image, and obtaining a deviation of the vehicle relative to the lane line based on the lane line semantic information; The vehicle posture information is normalized using the deviation of the vehicle relative to the lane line.

3. The vehicle positioning method in a multi-layer space transition area according to claim 1, characterized in that: After collecting the vehicle location information, the method further includes: Structural information constraints of a multi-layer spatial transition region are obtained, and the vehicle position error is optimized based on the structural information constraints of the multi-layer spatial transition region.

4. The vehicle positioning method in a multi-layer space transition area according to claim 1, characterized in that: The step of associating the vehicle position information with the vehicle posture information, the bird's-eye view image information, and the altitude information to construct a posture feature map for the multi-layer spatial transition area includes: Traverse the multi-layer spatial transition area and convert the collected data into vectors Organizing and storing, completing the construction of the posture feature map for the multi-layer space transition area, wherein: x (i) , y (i) For the vehicle i The vehicle location information corresponding to each node, a (i) , r (i) , p (i) For the vehicle i The pose information at each node, h (i) For the vehicle i The altitude information at each node, I (i) For the collection of i The bird's-eye view image information of each node, is the number of nodes.

5. The vehicle positioning method in a multi-layer space transition area according to claim 1, characterized in that: Before matching the pose feature map using the DTW algorithm to determine the pose feature candidate set, the method further includes: Construct a short-term uniform motion model based on the vehicle's historical position information; Determine a preliminary matching range of the vehicle in the posture feature map based on the short-term uniform motion model; The DTW algorithm is used to match the pose feature map to determine the pose feature candidate set, specifically including: The DTW algorithm is used to match the preliminary matching range in the pose feature map to determine the pose feature candidate set.

6. The vehicle positioning method in a multi-layer space transition area according to claim 5, characterized in that: The method takes the nodes in the pose feature candidate set as particles, takes the real-time vehicle pose feature as observation information, calculates the likelihood probability between the observation information and the particles, and outputs the best matching node in the pose feature candidate set, including: The likelihood probability between the real-time vehicle posture feature and each node in the posture feature candidate set is calculated, and the particles are sampled by importance to output the best matching node in the posture feature candidate set.

7. The vehicle positioning method in a multi-layer space transition area according to claim 6, characterized in that: The step of calculating the position and posture of the vehicle in the global coordinate system based on the best matching node includes: The relative position relationship between the real-time vehicle posture feature and the posture feature of each node in the best matching node is calculated, and the posture of the vehicle in the global coordinate system is solved through a preset coordinate transformation relationship.

8. A vehicle positioning device in a multi-layer space transition area, characterized in that: include: An information collection module is used to discretize the multi-layer spatial transition area into sequence nodes and collect the vehicle posture information, bird's-eye view image information and altitude information corresponding to each node; A map construction module, used for collecting vehicle position information, associating the vehicle position information with the vehicle posture information, the bird's-eye view image information and the altitude information, and constructing a posture feature map for the multi-layer space transition area; A matching module is used to match the pose feature map using the DTW algorithm to determine a candidate set of pose features; A calculation module, used to use the nodes in the pose feature candidate set as particles and the real-time vehicle pose feature as observation information, calculate the likelihood probability between the observation information and the particles, and output the best matching node in the pose feature candidate set; The vehicle positioning module is used to calculate the position and posture of the vehicle in the global coordinate system based on the best matching node.

9. An electronic device, characterized in that: comprising a memory and a processor, wherein: The memory is used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the vehicle positioning method in a multi-layer spatial transition area as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the vehicle positioning method in a multi-layer spatial transition area as described in any one of claims 1 to 7.

Images