Automatic parking method, device and equipment based on composite map and storage medium

Abstract

The invention relates to the technical field of automatic parking, in particular to an automatic parking method and device based on a composite map, equipment and a storage medium, and the method comprises the steps: firstly obtaining a feature point cloud map and a semantic map in a current environment through a real-time environment image, and then fusing the feature point cloud map and the semantic map to obtain a composite parking map with various attributes; and according to the composite parking map, a target parking space and a parking route are determined, loopback control is completed through a track map, and it is ensured that the vehicle can reach the area where the reverse parking operation is executed according to a preset parking path so that the automatic parking task can be completed. According to the scheme, the composite map with multiple attributes is used as the basis of route planning and vehicle control, the adaptive capacity and parking precision of the automatic parking system to a complex parking environment are remarkably improved, the recognition capacity of the system to obstacles and empty parking spaces is enhanced, the vehicle can flexibly cope with emergencies, and the safety of the vehicle is improved. And the safety and the efficiency in the parking process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automatic parking, and in particular to an automatic parking method, device, equipment and storage medium based on a composite map. Background Art

[0002] With the rapid development of science and technology, autonomous driving technology has become an important development direction of the automotive industry. The core of autonomous driving technology lies in the three major systems of environmental perception, decision planning and control execution. Among them, the environmental perception system is responsible for real-time monitoring of the environment around the vehicle, including obstacles, traffic signs, other vehicles and pedestrians. At present, autonomous driving technology still faces many challenges in practical applications, especially in complex urban environments and parking scenarios. How to accurately perceive the environment and make correct decisions has become the key to technological development.

[0003] As an important branch of autonomous driving technology, automatic parking aims to enable vehicles to automatically identify parking spaces and complete parking operations autonomously. This technology is of great significance in improving the utilization rate of parking lots, reducing the time spent looking for parking spaces, and alleviating traffic congestion. However, automatic parking technology faces a series of challenges in practical applications. First, the parking environment is complex and changeable, including different lighting conditions, dynamically changing obstacles, and interference from other vehicles. Second, parking spaces vary in size and shape, requiring the system to accurately identify parking space boundaries and assess whether the parking space is suitable for parking. In addition, planning a safe and efficient parking path is also one of the technical difficulties. These challenges require that the automatic parking system must have high-precision environmental perception capabilities and intelligent decision-making and planning capabilities.

[0004] Therefore, how to improve the recognition accuracy of environmental elements of the automatic parking system in complex parking environments, so as to accurately identify parking space boundaries, evaluate parking space availability, and effectively plan parking routes, has become a technical problem that needs to be urgently solved in this industry. Summary of the invention

[0005] The main purpose of the present invention is to provide an automatic parking method, device, equipment and storage medium based on a composite map, aiming to solve the technical problem in the prior art of how to improve the recognition accuracy of environmental elements of the automatic parking system in a complex parking environment, so as to accurately identify the boundaries of parking spaces, evaluate the availability of parking spaces, and effectively plan parking paths.

[0006] To achieve the above object, the present invention provides an automatic parking method based on a composite map, the method comprising the following steps: In response to the automatic parking instruction, obtaining a composite parking map under the current environment; Determining a target parking space and a parking route according to the composite parking map; The vehicle is controlled to drive into the target parking space according to the parking route to complete the automatic parking task.

[0007] Optionally, the obtaining of the composite parking map in the current environment includes: In response to an automatic parking instruction, acquiring a real-time environment image; According to the real-time environment image, a feature point cloud map and a semantic map of the current environment are obtained; The feature point cloud map is fused with the semantic map to obtain the composite parking map.

[0008] Optionally, obtaining a feature point cloud map in the current environment according to the real-time environment image includes: Acquire a forward-looking monocular image in the real-time environment image; Using the front-view monocular image as input data of a feature point detection model to obtain a visual feature point extraction result; Based on the vehicle posture data, inter-frame matching is performed on the visual feature point extraction results to obtain inter-frame feature point pairs; The feature point cloud map is obtained according to the inter-frame feature point pairs.

[0009] Optionally, obtaining a semantic map of the current environment according to the real-time environment image includes: Acquire four-surround view images in the real-time environment image; Obtaining a bird's-eye view map according to the four surround view images; Performing semantic feature recognition on each block of the bird's-eye view map to obtain a semantic feature recognition result, wherein the semantic feature includes parking sign features and obstacle features; The semantic map is obtained by performing semantic feature recognition results on each block of the bird's-eye view map.

[0010] Optionally, determining a target parking space according to the composite parking map includes: Determining a parking space to be selected according to the semantic map in the composite parking map; Obtaining the parking space of the to-be-selected parking space according to the feature point cloud map in the composite parking map; Determining a parking risk assessment score corresponding to the to-be-selected parking space according to the parking space; A target parking space among the to-be-selected parking spaces is determined according to the parking risk assessment score.

[0011] Optionally, obtaining the storage space of the parking space to be selected according to the feature point cloud map in the composite parking map includes: Obtaining adjacent area information and adjacent lane information of the to-be-selected parking space according to the feature point cloud map; According to the adjacent area information, a vehicle movement space in the warehouse is obtained; According to the lane information, a vehicle movement space outside the warehouse is obtained; The storage space is obtained according to the vehicle movement space within the storage and the vehicle movement space outside the storage.

[0012] Optionally, determining a parking route according to the composite parking map includes: Determine a route to be selected based on the real-time position of the vehicle and the target parking space; Obtaining obstacle information and vehicle information within a preset area of ​​the to-be-selected route according to the semantic map in the composite parking map; A parking route among the candidate routes is determined according to the obstacle information and the vehicle information.

[0013] In addition, to achieve the above-mentioned purpose, the present invention further proposes an automatic parking device based on a composite map, the automatic parking device based on a composite map comprising: A map construction module, for obtaining a composite parking map under a current environment in response to an automatic parking instruction; A parking planning module, used to determine a target parking space and a parking route according to the composite parking map; The vehicle control module is used to control the vehicle to enter the target parking space according to the parking route to complete the automatic parking task.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also proposes an automatic parking device based on a composite map, and the automatic parking device based on a composite map includes: a memory, a processor, and an automatic parking program based on a composite map stored in the memory and executable on the processor, and the automatic parking program based on a composite map is configured to implement the steps of the automatic parking method based on a composite map as described above.

[0015] In addition, to achieve the above-mentioned purpose, the present invention also proposes a storage medium, on which an automatic parking program based on a composite map is stored. When the automatic parking program based on a composite map is executed by a processor, the steps of the automatic parking method based on a composite map as described above are implemented.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: This solution uses a composite map with multiple navigation attributes as the basis for route planning and vehicle control, which significantly improves the adaptability and parking accuracy of the automatic parking system to complex parking environments, enhances the system's ability to identify obstacles and empty parking spaces, enables the vehicle to respond flexibly to emergencies, and improves the safety and efficiency of the parking process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0019] Figure 1 It is a flowchart of a first embodiment of an automatic parking method based on a composite map of the present invention; Figure 2 It is a flow chart of a second embodiment of the automatic parking method based on a composite map of the present invention; Figure 3 It is a structural block diagram of a first embodiment of an automatic parking device based on a composite map of the present invention; Figure 4 It is a structural diagram of an automatic parking device based on a composite map in a hardware operating environment involved in an embodiment of the present invention.

[0020] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0021] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application and are not used to limit the present application.

[0022] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of the embodiment of the present application is to achieve accurate perception of the current environment by constructing a composite map based on a feature point cloud map and a semantic map, thereby determining the target parking space and planning the parking route. Using this composite map, the system can evaluate the storage space and parking risk of the selected parking space, determine the best parking space based on the evaluation results, and plan the parking path under the target parking space. Finally, the loop control is completed through the trajectory map, so that the vehicle drives into the target parking space along the planned path and completes the automatic parking task.

[0024] In the existing automatic parking process, the system first needs to perceive obstacles, empty parking spaces and other vehicles in the surrounding environment, and then find available parking spaces. This process requires accurate identification of the boundaries of available parking spaces and evaluation of whether there are enough parking spaces for vehicles. After identifying the parking spaces, a path from the current position to the parking space needs to be planned. This process also requires identification of various environmental elements on the path. Therefore, the overall focus is on improving the recognition accuracy of environmental elements, thereby obtaining an accurate parking scene map and completing the parking task.

[0025] Therefore, how to improve the recognition accuracy of environmental elements of the automatic parking system in complex parking environments, so as to accurately identify parking space boundaries, evaluate parking space availability, and effectively plan parking routes, is a technical problem that urgently needs to be solved.

[0026] This application uses a composite map with multiple attributes as the basis for route planning and vehicle control, which significantly improves the adaptability and parking accuracy of the automatic parking system to complex parking environments, enhances the system's ability to identify obstacles and empty parking spaces, enables the vehicle to respond flexibly to emergencies, and improves the safety and efficiency of the parking process.

[0027] It should be noted that the execution subject of the present invention may be an automatic parking device based on a composite map, or a computing service device with data processing, network communication and program running functions, such as a tablet computer, a personal computer, a mobile phone, etc., or an automatic parking device based on a composite map that can realize the above functions, etc., and this embodiment does not specifically limit this. The following takes the automatic parking device based on a composite map as an example of the execution subject to illustrate this embodiment and the following embodiments.

[0028] Based on this, an embodiment of the present invention provides an automatic parking method based on a composite map, referring to Figure 1 , Figure 1 The figure is a flow chart of a first embodiment of an automatic parking method based on a composite map according to the present invention.

[0029] In this embodiment, the automatic parking method based on the composite map includes: Step S10: In response to the automatic parking instruction, a composite parking map under the current environment is obtained.

[0030] It should be noted that when the vehicle is driving normally, the vehicle system can automatically generate an obstacle avoidance map through its own external perception unit. The principle is to collect data on the surrounding environment through onboard sensors such as lidar, ultrasonic radar, stereo cameras, etc., and parse out information such as the position, speed and shape of obstacles around the vehicle from these data, and then generate information to guide the vehicle in obstacle avoidance and path planning.

[0031] It is understandable that conventional map construction is limited to a few types of maps, that is, only one of the point cloud map, semantic map, and grid map is generated. It is rare to generate multiple maps with different attributes for path planning at the same time. Different types of maps have their own advantages, which are specifically manifested in different tasks of vehicle autonomous driving. For example, the point cloud map can retain the original environmental measurement information in the model as much as possible, and has better effects in the process of precise positioning and obstacle detection; the grid map can represent the occupancy status of obstacles, and has better effects in path planning for short paths; the feature map represents the environment through geometric features (such as points, lines, and surfaces), with relatively small data storage and computing volume, and is suitable for resource-constrained systems; the semantic map can better understand the deep essence of the current environment, and can provide accurate identification results of various elements in the road, such as the meaning of traffic signs, the type of road markings, etc., so that the autonomous driving system can help the autonomous driving system make more reasonable driving decisions, such as how to choose the correct driving path at a complex traffic intersection. It can be seen that any map with a single attribute cannot effectively meet the needs of the autonomous driving system in all scenarios. Therefore, in order to overcome the limitations of a single map type, the field of autonomous driving is increasingly inclined to adopt a multi-map fusion strategy, that is, to simultaneously generate and use multiple maps with different attributes to improve the overall performance of the system. This fusion method can combine the advantages of each map to provide more comprehensive and accurate environmental information, thereby improving the adaptability to complex scenarios.

[0032] It should be understood that the composite map in the present invention is only used to construct relevant maps based on environmental factors in automatic parking scenarios. Specifically, when the driver enters the parking lot or intends to park, the vehicle automatically activates the construction of the composite parking map and the subsequent automatic parking function through relevant automatic parking instructions.

[0033] In one embodiment, the step of obtaining a composite parking map in the current environment in response to an automatic parking instruction includes: obtaining a real-time environment image in response to an automatic parking instruction; obtaining a feature point cloud map and a semantic map in the current environment based on the real-time environment image; and performing map fusion on the feature point cloud map and the semantic map to obtain the composite parking map.

[0034] It should be noted that the real-time environmental images include but are not limited to the forward-looking monocular image and the four-surround view image. The forward-looking monocular image is mainly used for detecting the road ahead and obstacles, while the four-surround view image provides a panoramic view around the vehicle, which helps to fully understand the environment around the vehicle.

[0035] Furthermore, obtaining a feature point cloud map in the current environment based on the real-time environment image includes: acquiring a forward-looking monocular image in the real-time environment image; using the forward-looking monocular image as input data of a feature point detection model to obtain a visual feature point extraction result; performing inter-frame matching on the visual feature point extraction result based on vehicle posture data to obtain inter-frame feature point pairs; and obtaining the feature point cloud map based on the inter-frame feature point pairs.

[0036] It should be noted that the forward-looking monocular image refers to the two-dimensional image captured by the vehicle's forward camera, which provides visual information of the environment in front of the vehicle. By analyzing these images, the system can identify key features such as roads, obstacles, and traffic signs. The feature point detection model is a computer vision algorithm used to extract key points (feature points) from images. These feature points are usually some significant and unique points in the image, such as corner points, edge points or spots. They have good stability and recognizability under different viewing angles or lighting conditions; inter-frame matching refers to the process of finding the same feature points between consecutive image frames. By comparing the feature points in consecutive frames, the motion trajectory of the feature points in space can be determined. Combined with the vehicle's posture information, the change information of the feature points in space and time can be obtained, thereby constructing a three-dimensional point cloud, and then obtaining a feature point cloud map, which records in detail the three-dimensional spatial information within a certain scale around the vehicle.

[0037] It should be understood that, based on the characteristics of point cloud maps, point cloud maps are more suitable for performing tasks related to precise vehicle control, such as controlling the vehicle to maintain a distance from vehicles in adjacent parking spaces during automatic reversing into a garage, or controlling the vehicle body angle when meeting other vehicles in a narrow lane.

[0038] Furthermore, obtaining a semantic map in the current environment based on the real-time environment image includes: acquiring a four-surround view image in the real-time environment image; obtaining a bird's-eye view map based on the four-surround view image; performing semantic feature recognition on each block of the bird's-eye view map to obtain a semantic feature recognition result, wherein the semantic feature includes parking sign features and obstacle features; and performing semantic feature recognition results on each block of the bird's-eye view map to obtain the semantic map.

[0039] It can be understood that four-way view images refer to images collected from all around the vehicle. These images can provide a 360-degree view of the environment around the vehicle, which is crucial for building a bird's-eye view. The bird's-eye view provides the vehicle with comprehensive spatial perception capabilities and provides a perspective for observing the vehicle's surroundings from above. Since vehicle route planning is essentially a two-dimensional shortest path planning, this perspective helps to identify parking spaces, obstacles and other important traffic elements, and can simplify complex road structures, making them easier to analyze and understand. Semantic feature recognition refers to identifying features with specific meanings from a bird's-eye view, such as parking sign features (such as parking spaces, lane lines) and obstacle features (such as vehicles, pedestrians, buildings). This step is to identify and classify different objects and signs in the image through a deep learning algorithm, and then fill the recognition results into the map according to the original position to form a semantic map. The semantic map contains not only geometric information, but also rich annotation information, such as the vectorized representation of the road, type (lane, bicycle lane or bus lane), speed limit, lane line type, etc.

[0040] It should be understood that, based on the characteristics of semantic maps, semantic maps are more suitable for tasks that require understanding the deep meaning of the environment and traffic rules. For example, in the process of automatic parking, semantic maps can help vehicles identify key information such as parking space boundaries, no-parking areas, traffic signs, etc., so as to make decisions on whether it is suitable to park.

[0041] Step S20: Determine a target parking space and a parking route according to the composite parking map.

[0042] It should be noted that since the composite parking map includes both the point cloud map and the semantic map, the point cloud map provides precise distance and shape information, while the semantic map provides semantic information about parking signs, traffic rules, etc. Based on the constructed map resources, the available parking spaces can be traversed. When traversing to the free parking spaces in the current area, the system will trigger the path planning algorithm to calculate the best path from the current position to the target parking space. When everything is ready, the system will automatically perform the parking operation and park the vehicle in the determined free parking space.

[0043] It is understandable that the system needs to update the information of vacant parking spaces in real time to reflect status changes that may be caused by other vehicles parking or leaving. For example, other vehicles may have parked in the original vacant parking space before this vehicle arrives. At this time, it is necessary to select a new target parking space. In some cases, the system may detect multiple vacant parking spaces. At this time, it is necessary to determine the most suitable vacant parking space based on the preset parking space selection logic, such as evaluating the size, location, distance from the destination, parking difficulty and other factors of the multiple optional parking spaces to determine the final parking space.

[0044] It should be understood that after determining the target parking space, the system needs to plan a path from the current position to the target parking space. The planning process also needs to take into account the dynamic impact of obstacles, traffic rules and other vehicles in the composite parking map. Since the composite parking map provides accurate distance and shape information (point cloud map) and semantic information about parking signs, traffic rules, etc. (semantic map), the system can better comply with traffic rules and avoid conflicts with obstacles and other vehicles, thereby improving the safety and efficiency of the parking process.

[0045] Step S30: Control the vehicle to drive into the target parking space according to the parking route to complete the automatic parking task.

[0046] It should be noted that, based on the aforementioned composite parking map, according to all turning points and intersections during driving, the system can generate a corresponding trajectory map, thereby predicting and planning the vehicle's driving trajectory to avoid collisions in areas where detection may fail, such as turns or intersections.

[0047] It can be understood that the trajectory map adds a time dimension and displays the geographic location data of different time nodes in the same chart, so that users can intuitively see the position and path of objects (such as vehicles, animals, etc.) changing over time. In the automatic parking system, the trajectory map can help vehicles drive from the entrance of the parking lot to the vicinity of the parking space, involving path planning for the driving and parking stages. The trajectory map can record users' frequently used parking locations and parking trajectories, construct a feature map of the parking environment, and realize autonomous learning of parking behavior.

[0048] It should be understood that the track map not only records the historical parking track, but also updates the current position and motion status of the vehicle in real time. This enables the system to dynamically adjust the parking path according to the latest environmental changes and vehicle status to deal with emergencies, such as the sudden insertion of other vehicles or the crossing of pedestrians. In complex or changing environments, such as complex structures or insufficient light in parking lots, the track map can help the system maintain robustness. Even in areas where sensor detection may fail, the system can continue to navigate based on historical data in the track map.

[0049] This embodiment achieves comprehensive perception of the vehicle's surroundings by fusing point cloud maps and semantic maps with real-time environmental images. In the solution, the system first responds to the automatic parking command, obtains the real-time environmental image, and extracts the feature point cloud map and semantic map from it; then, the two maps are fused to form a composite parking map. Using the composite parking map, the system can accurately identify parking spaces, obstacles and traffic signs, plan the best path from the vehicle's current position to the target parking space, and control the vehicle to automatically drive into the parking space to complete the parking task. In addition, the system can also generate trajectory maps based on turning points and intersections during driving, predict and plan the vehicle's driving trajectory, and avoid collisions in areas where detection may fail.

[0050] In summary, this solution significantly improves the environmental perception capability and path planning accuracy of the automatic parking system. The fused map provides more comprehensive environmental information, enabling the system to better comply with traffic rules and avoid conflicts with obstacles and other vehicles, thereby improving the safety and efficiency of the parking process. The introduction of the trajectory map enables the system to update the vehicle position and motion status in real time, dynamically adjust the parking path to respond to emergencies, enhance the robustness of the system, and provide users with more personalized and convenient parking services.

[0051] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction, and will not be repeated in the following. Figure 2 , the step S20 comprises: Step S201: determining a parking space to be selected according to the semantic map in the composite parking map.

[0052] It should be noted that since there is a bird's-eye view mode in the display format of the semantic map, in the bird's-eye view mode, you can first traverse whether there is information about vacant parking spaces in the current map. If the traversal cannot be found or is beyond the perception range, you need to control the vehicle to perceive the information of the next area and find vacant parking spaces. At this time, you need to control the vehicle to continue driving along the current lane or drive to the next unknown area until you traverse to an available parking space.

[0053] It is understandable that the number of parking spaces to be selected will generally change dynamically during the driving process of the vehicle, and the system will evaluate the safety and convenience of each parking space in real time. For example, parking path planning may also need to consider the vehicle's dynamic limitations, such as the minimum turning radius, acceleration and deceleration limits, etc. The more maneuvering space is reserved in the parking space to be selected, the more convenient parking will be. In addition, the number of obstacles that require extra attention, such as vehicles and pedestrians near the route to the target parking space, can further determine the safety of parking.

[0054] Step S202: obtaining the parking space of the parking space to be selected according to the feature point cloud map in the composite parking map.

[0055] It should be noted that the parking process from the current position to the parking space can be divided into two stages. The first stage is to move from the current position to the area where the entry operation is to be performed. Before reaching the area where the entry operation is to be performed, the vehicle completes loop control through the trajectory map, so that the vehicle arrives at the area where the entry operation is to be performed according to the parking path. During the process, the position of the vehicle is corrected through real-time environmental images, and the parking route and the target parking space need to be corrected when necessary. The second stage is in the area where the entry operation can be performed. After reaching the position, the system will obtain the feature point cloud map within a certain spatial range, and then perform the preset reverse entry operation. The loop control is performed through the combination of the feature point cloud map and the semantic map, thereby completing the automatic parking task.

[0056] It is understandable that since the vehicle's reversing operation into a parking space is a relatively fixed vehicle movement operation, the vehicle can be accurately parked in the parking space by following the preset reversing control procedure when the vehicle reaches a certain area in the left front or right front of the parking space. Therefore, at this stage, more attention should be paid to the relative position between the vehicle and obstacles in the adjacent range, that is, parking control should be performed according to the area within a certain range of the parking space.

[0057] In one embodiment, based on the feature point cloud map, adjacent area information and adjacent lane information of the selected parking space are obtained; based on the adjacent area information, the vehicle movement space within the garage is obtained; based on the lane information, the vehicle movement space outside the garage is obtained; based on the vehicle movement space within the garage and the vehicle movement space outside the garage, the storage space is obtained.

[0058] It is understandable that the adjacent area information refers to the relevant information in the parking spaces on the left and right sides of the parking space line of the selected parking space, including the distance to the wall, the location of obstacles, the pedestrian area, etc. Based on this, the movable range of the vehicle body after entering the garage is obtained as the vehicle movement space in the garage. Since the vehicle needs to have enough space for turning and reversing when entering the garage, the specific location of fixed or mobile obstacles in the lanes adjacent to the parking spaces, such as other vehicles, pillars, trash cans, etc., is also required on the basis of the vehicle movement space in the garage. This information is very necessary for planning parking paths and avoiding collisions. They are used to determine the movable range of the vehicle after entering the garage, that is, the vehicle movement space in the garage. In addition, various traffic signs and rules on the lanes, such as speed limit signs, parking signs and other information, also help the system comply with traffic rules and plan parking operations that comply with regulations.

[0059] Step S203: determining a parking risk assessment score corresponding to the parking space to be selected according to the storage space.

[0060] It should be noted that the assessment of parking risks includes two aspects: on the one hand, the assessment of the route to the parking space, specifically including the total length of the assessment route. The shorter the route means less maneuvering and lower risk, or whether there are visual obstructions on the assessment route, such as buildings, vehicles, etc., which may affect the driver's response to emergencies. It also includes the assessment of the number of pedestrians and vehicles on the route. The fewer the number, the lower the risk. On the other hand, there is the assessment of vehicles on the left and right of the parking space. If there are no cars on both sides of the parking space, this provides maximum flexibility and minimum risk. When there is a car on only one side, it is necessary to assess whether the space between the two vehicles is sufficient for the vehicles to enter and exit safely. If there are cars on both sides, the width of the space left by the two cars is assessed. The wider the space, the lower the difficulty of parking. Furthermore, the size of adjacent vehicles is also considered. Large vehicles may take up more space and affect parking. The parking angle of adjacent vehicles and irregular parking will also affect the assessment results.

[0061] It is understandable that a scoring system can be designed based on the above-mentioned situations, quantifying the above-mentioned factors into scores, assigning different weights to each factor according to its impact on parking difficulty and risk, and judging the suitability of the parking space based on the total score. The higher the score, the lower the parking risk and the more suitable the parking space.

[0062] Step S204: determining a target parking space among the to-be-selected parking spaces according to the parking risk assessment score.

[0063] It should be noted that the parking risk assessment score is a comprehensive indicator that reflects the safety and feasibility of each parking space to be selected. This score is based on real-time analysis of the surrounding environment, including obstacles, traffic flow, parking space size and other factors. It can be used to sort and screen the parking spaces to be selected. The system will compare the risk scores of all parking spaces to be selected and select the parking space with the lowest risk as the target parking space to ensure the safety and efficiency of the parking process.

[0064] It is understandable that the parking risk assessment score is only one criterion in the decision-making process. In addition, the system needs to consider other factors, such as the relative distance of the parking space to the exit, user preference, parking costs, etc., which may be taken into account together with the risk assessment score to make the final parking decision.

[0065] It should be understood that when selecting a target parking space, the system needs to find a balance between risk and convenience. For example, a parking space with a lower risk may be farther from the destination, while a parking space with a slightly higher risk may be closer to the destination. The system needs to weigh these factors based on the user's needs and preferences.

[0066] Step S205: Determine a route to be selected based on the real-time position of the vehicle and the target parking space.

[0067] It should be noted that the vehicle's real-time location data is obtained through on-board GPS, on-board sensors (such as lidar, cameras) and inertial measurement units (IMUs) and other devices. These data are crucial in determining the vehicle's current position and orientation, and are the basis for planning alternative routes.

[0068] It is understandable that through the accurate positioning and road network information of the semantic map, the autonomous driving system can determine the best path from the current position to the target parking space based on the road and traffic information in the semantic map. When planning the candidate route, the system needs to consider the road and traffic information in the semantic map, such as lanes, bicycle lanes or bus lanes), speed limits, lane line types, etc., to ensure that the planned route is not only technically feasible, but also conforms to the actual traffic logic.

[0069] Step S206: Obtaining obstacle information and vehicle information within the preset area of ​​the to-be-selected route according to the semantic map in the composite parking map.

[0070] It should be noted that obstacle information refers to the size, shape, position of obstacles on the selected route, as well as the type, speed and direction of travel of vehicles. The system needs to identify fixed obstacles such as walls, pillars, other vehicles, etc., as well as possible dynamic obstacles such as pedestrians, bicycles and other moving vehicles. In particular, the vehicle information in the route needs to be evaluated separately.

[0071] It is understandable that vehicle information includes the type, speed and driving direction of each vehicle on the route, because the vehicle is the most direct interactive object in the autonomous driving process, and its dynamic characteristics have a direct impact on path planning and obstacle avoidance strategies.

[0072] Step S207: Determine a parking route among the candidate routes according to the obstacle information and the vehicle information.

[0073] It should be noted that when planning parking routes, the system will give priority to safety. Even if a route is technically feasible, if there is a high safety risk, the system will choose other safer routes. The risk assessment mainly includes the number of moving obstacles on the route, that is, the number of pedestrians and bicycles, and the number of encounters caused by oncoming vehicles. Therefore, the system needs to balance multiple goals, such as the shortest path, minimum obstacles, minimum energy consumption and maximum comfort. The system needs to combine these factors to find the best parking route.

[0074] This embodiment realizes automatic parking of autonomous vehicles by combining semantic maps and feature point cloud maps in the composite parking map. First, the system determines the parking space to be selected and evaluates its safety and convenience; then, it obtains the storage space of the parking space to be selected and determines the parking risk assessment score based on this space; then, it selects the target parking space based on the risk assessment score and plans the best route from the current position to the target parking space; finally, the system obtains obstacle and vehicle information on the route to determine the final parking route. The whole process involves real-time environmental perception, path planning, risk assessment and dynamic obstacle avoidance to ensure the safety and efficiency of the parking process.

[0075] In summary, the technical solution of this embodiment improves the accuracy and safety of automatic parking of autonomous vehicles by comprehensively utilizing semantic maps and feature point cloud maps. By evaluating the safety and convenience of parking spaces in real time, the system can dynamically adjust parking strategies and select the best parking space. At the same time, through accurate storage space analysis and risk assessment, the system can reduce potential risks in the parking process and improve parking efficiency.

[0076] This application also provides an automatic parking device based on a composite map, please refer to Figure 3 , the automatic parking device based on the composite map includes: A map construction module 10, for obtaining a composite parking map under a current environment in response to an automatic parking instruction; A parking planning module 20, for determining a target parking space and a parking route according to the composite parking map; The vehicle control module 30 is used to control the vehicle to drive into the target parking space according to the parking route to complete the automatic parking task.

[0077] In one embodiment, the map construction module 10 is further used to obtain a real-time environment image in response to an automatic parking instruction; obtain a feature point cloud map and a semantic map in the current environment based on the real-time environment image; and perform map fusion on the feature point cloud map and the semantic map to obtain the composite parking map.

[0078] In one embodiment, the map construction module 10 is also used to obtain a forward-looking monocular image in the real-time environment image; use the forward-looking monocular image as input data of a feature point detection model to obtain a visual feature point extraction result; perform inter-frame matching on the visual feature point extraction result based on vehicle posture data to obtain inter-frame feature point pairs; and obtain the feature point cloud map based on the inter-frame feature point pairs.

[0079] In one embodiment, the map construction module 10 is further used to obtain a four-surround view image in the real-time environment image; obtain a bird's-eye view map based on the four-surround view image; perform semantic feature recognition on each block of the bird's-eye view map to obtain a semantic feature recognition result, wherein the semantic feature includes parking sign features and obstacle features; perform semantic feature recognition results on each block of the bird's-eye view map to obtain the semantic map.

[0080] In one embodiment, the parking planning module 20 is further used to determine a parking space to be selected based on the semantic map in the composite parking map; obtain a storage space of the parking space to be selected based on the feature point cloud map in the composite parking map; determine a parking risk assessment score corresponding to the parking space to be selected based on the storage space; and determine a target parking space among the parking spaces to be selected based on the parking risk assessment score.

[0081] In one embodiment, the parking planning module 20 is further used to obtain the adjacent area information and adjacent lane information of the selected parking space based on the feature point cloud map; obtain the vehicle movement space within the garage based on the adjacent area information; obtain the vehicle movement space outside the garage based on the lane information; and obtain the parking space based on the vehicle movement space within the garage and the vehicle movement space outside the garage.

[0082] In one embodiment, the parking planning module 20 is further used to determine a candidate route based on the real-time position of the vehicle and the target parking space; obtain obstacle information and vehicle information within a preset area of ​​the candidate route based on a semantic map in the composite parking map; and determine a parking route in the candidate route based on the obstacle information and vehicle information.

[0083] This embodiment uses a composite map with multiple attributes as the basis for route planning and vehicle control, which significantly improves the adaptability and parking accuracy of the automatic parking system to complex parking environments, enhances the system's ability to identify obstacles and empty parking spaces, enables the vehicle to flexibly respond to emergencies, and improves the safety and efficiency of the parking process.

[0084] The composite map-based automatic parking device provided by the present application adopts the composite map-based automatic parking method in the above embodiment, which can solve the technical problem of how to improve the recognition accuracy of environmental elements of the automatic parking system in a complex parking environment, so as to accurately identify the parking space boundary, evaluate the parking space availability, and effectively plan the parking path. Compared with the prior art, the beneficial effects of the composite map-based automatic parking device provided by the present application are the same as the beneficial effects of the composite map-based automatic parking method provided by the above embodiment, and the other technical features of the composite map-based automatic parking device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0085] The present application provides an automatic parking device based on a composite map, and the automatic parking device based on a composite map includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the automatic parking method based on the composite map in the above-mentioned embodiment one.

[0086] Reference below Figure 4 , which shows a schematic diagram of the structure of the automatic parking device based on the composite map suitable for implementing the embodiment of the present application. The automatic parking device based on the composite map in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The composite map-based automatic parking device shown is merely an example and should not impose any limitation on the functions and scope of use of the embodiments of the present application.

[0087] like Figure 4As shown, the automatic parking device based on the composite map may include a processing device 1001 (such as a central processing unit, a graphics processor, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 to a random access memory (RAM: Random Access Memory) 1004. In RAM1004, various programs and data required for the operation of the automatic parking device based on the composite map are also stored. The processing device 1001, ROM1002 and RAM1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems may be connected to the I / O interface 1006: input devices 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; storage devices 1003 including, for example, a magnetic tape, a hard disk, etc.; and communication devices 1009. The communication device 1009 may allow the composite map-based automatic parking device to communicate with other devices wirelessly or by wire to exchange data. Although the composite map-based automatic parking device with various systems is shown in the figure, it should be understood that it is not required to implement or have all the systems shown. More or fewer systems may be implemented or have alternatively.

[0088] In particular, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are executed.

[0089] The composite map-based automatic parking device provided by the present application adopts the composite map-based automatic parking method in the above embodiment, which can solve the technical problem of how to improve the recognition accuracy of environmental elements of the automatic parking system in a complex parking environment, so as to accurately identify the parking space boundary, evaluate the parking space availability, and effectively plan the parking path. Compared with the prior art, the beneficial effects of the composite map-based automatic parking device provided by the present application are the same as the beneficial effects of the composite map-based automatic parking method provided by the above embodiment, and the other technical features of the composite map-based automatic parking device are the same as the features disclosed in the method of the previous embodiment, which will not be repeated here.

[0090] It should be understood that the various parts disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0091] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0092] The present application provides a computer-readable storage medium having computer-readable program instructions (ie, computer programs) stored thereon, the computer-readable program instructions being used to execute the composite map-based automatic parking method in the above-mentioned embodiment.

[0093] The computer-readable storage medium provided in the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, systems or devices, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, system or device. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: Radio Frequency), etc., or any suitable combination of the above.

[0094] The computer-readable storage medium may be included in the automatic parking device based on the composite map; or may exist independently without being assembled into the automatic parking device based on the composite map.

[0095] The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the automatic parking device based on the composite map, the automatic parking device based on the composite map: responds to the automatic parking instruction, obtains the composite parking map in the current environment; determines the target parking space and the parking route according to the composite parking map; controls the vehicle to enter the target parking space according to the parking route, so as to complete the automatic parking task.

[0096] Computer program code for performing the operations of the present application may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0097] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0098] The modules involved in the embodiments described in this application may be implemented by software or hardware, wherein the name of the module does not constitute a limitation on the unit itself in some cases.

[0099] The readable storage medium provided in this application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned automatic parking method based on a composite map, and can solve the technical problem of how to improve the recognition accuracy of environmental elements of the automatic parking system in a complex parking environment, so as to accurately identify the boundaries of parking spaces, evaluate the availability of parking spaces, and effectively plan parking paths. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the automatic parking method based on a composite map provided in the above-mentioned embodiment, and will not be repeated here.

[0100] The computer program product provided by the present application can solve the technical problem of automatic parking based on composite maps. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as the beneficial effects of the automatic parking method based on composite maps provided by the above embodiment, which will not be repeated here.

[0101] The above descriptions are only some embodiments of the present application, and are not intended to limit the patent scope of the present application. All equivalent structural changes made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect applications in other related technical fields are included in the patent protection scope of the present application.

Claims

1. An automatic parking method based on a composite map, characterized in that: The composite map-based automatic parking method comprises: In response to the automatic parking instruction, obtaining a composite parking map under the current environment; Determining a target parking space and a parking route according to the composite parking map; The vehicle is controlled to drive into the target parking space according to the parking route to complete the automatic parking task.

2. The automatic parking method based on composite map according to claim 1, characterized in that: The step of obtaining a composite parking map under a current environment in response to the automatic parking instruction includes: In response to an automatic parking instruction, acquiring a real-time environment image; According to the real-time environment image, a feature point cloud map and a semantic map of the current environment are obtained; The feature point cloud map is fused with the semantic map to obtain the composite parking map.

3. The automatic parking method based on composite map according to claim 2, characterized in that: The step of obtaining a feature point cloud map in the current environment according to the real-time environment image includes: Acquire a forward-looking monocular image in the real-time environment image; Using the front-view monocular image as input data of a feature point detection model to obtain a visual feature point extraction result; Based on the vehicle posture data, inter-frame matching is performed on the visual feature point extraction results to obtain inter-frame feature point pairs; The feature point cloud map is obtained according to the inter-frame feature point pairs.

4. The automatic parking method based on composite map according to claim 2, characterized in that: The step of obtaining a semantic map of the current environment according to the real-time environment image includes: Acquire four-surround view images in the real-time environment image; Obtaining a bird's-eye view map according to the four surround view images; Performing semantic feature recognition on each block of the bird's-eye view map to obtain a semantic feature recognition result, wherein the semantic feature includes parking sign features and obstacle features; The semantic map is obtained by performing semantic feature recognition results on each block of the bird's-eye view map.

5. The automatic parking method based on composite map according to claim 1, characterized in that: Determining the target parking space according to the composite parking map includes: Determining a parking space to be selected according to the semantic map in the composite parking map; Obtaining the parking space of the to-be-selected parking space according to the feature point cloud map in the composite parking map; Determining a parking risk assessment score corresponding to the to-be-selected parking space according to the parking space; A target parking space among the to-be-selected parking spaces is determined according to the parking risk assessment score.

6. The automatic parking method based on composite map according to claim 5, characterized in that: The step of obtaining the parking space for the parking space to be selected according to the feature point cloud map in the composite parking map includes: Obtaining adjacent area information and adjacent lane information of the to-be-selected parking space according to the feature point cloud map; According to the adjacent area information, a vehicle movement space in the depot is obtained; According to the lane information, a vehicle movement space outside the warehouse is obtained; The storage space is obtained according to the vehicle movement space within the storage and the vehicle movement space outside the storage.

7. The automatic parking method based on composite map according to claim 1, characterized in that: Determining a parking route according to the composite parking map includes: Determine a route to be selected based on the real-time position of the vehicle and the target parking space; Obtaining obstacle information and vehicle information within a preset area of ​​the to-be-selected route according to the semantic map in the composite parking map; A parking route among the candidate routes is determined according to the obstacle information and the vehicle information.

8. An automatic parking device based on a composite map, characterized in that: The composite map-based automatic parking device comprises: A map construction module, for obtaining a composite parking map under a current environment in response to an automatic parking instruction; A parking planning module, used to determine a target parking space and a parking route according to the composite parking map; The vehicle control module is used to control the vehicle to enter the target parking space according to the parking route to complete the automatic parking task.

9. An automatic parking device based on a composite map, characterized in that: The composite map-based automatic parking device comprises: a memory, a processor, and a composite map-based automatic parking program stored in the memory and executable on the processor, wherein the composite map-based automatic parking program is configured to implement the steps of the composite map-based automatic parking method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores an automatic parking program based on a composite map, and when the automatic parking program based on a composite map is executed by a processor, the steps of the automatic parking method based on a composite map as described in any one of claims 1 to 7 are implemented.

Images