Map updating method, device and system and storage medium

Abstract

The embodiment of the invention provides a map updating method, device and system and a storage medium, and relates to the technical field of auxiliary driving. The method is applied to a vehicle end and comprises the steps that the vehicle end determines road surface obstacle information of a passing road section according to road condition point cloud; obtaining road surface addition and deletion information of the passing road section, wherein the road surface addition and deletion information of the passing road section is determined by the cloud according to a difference fusion result of the road surface information corresponding to the road condition point cloud and a cloud historical map; operation boundary addition and deletion information is obtained, the operation boundary addition and deletion information is determined by the roadside edge end according to the difference fusion result of the operation boundary update information and the historical map of the roadside edge end, and the operation boundary update information is determined by the roadside edge end processing vehicle end track data and the environment point cloud collected by the roadside edge end; and updating a vehicle end historical map according to the road surface obstacle information, the road surface addition and deletion information and the operation boundary addition and deletion information. According to the method, computing resources can be reasonably configured, the data transmission pressure is reduced, and timely and accurate map updating is realized.

Description

Technical Field

[0001] This application relates to the field of driver assistance technology, specifically to a map update method, device, system, and storage medium. Background Technology

[0002] Mining environments are characterized by high environmental dynamics, unstructured roads, and frequent changes in work areas. Autonomous vehicles rely on high-precision maps for accurate positioning, path planning, and obstacle avoidance. Traditional navigation maps cannot meet the centimeter-level accuracy and minute-level update requirements of mining areas, necessitating dynamic update technology to ensure timely map updates. Existing technologies primarily rely on professional surveying equipment, crowdsourced data collection, or SLAM mapping, but these methods struggle to achieve timely and accurate map updates in mining environments. Summary of the Invention

[0003] The purpose of this application is to provide a map updating method, apparatus, system, and storage medium.

[0004] To achieve the above objectives, the first aspect of this application provides a map updating method applied to a vehicle, the updating method comprising: Collect vehicle trajectory data and road condition point cloud data of the vehicle's travel route; Determine road obstacle information for the route based on road condition point cloud data; The information on road surface additions and deletions for the route is obtained. This information is determined by the cloud based on the differential fusion results of the road surface information corresponding to the road condition point cloud and the historical map in the cloud. The information on the addition and deletion of the operation boundary is obtained. The information on the addition and deletion of the operation boundary is determined by the roadside edge terminal based on the differential fusion result of the updated operation boundary information and the historical map of the roadside edge terminal. The updated operation boundary information is determined by the roadside edge terminal processing vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. The roadside edge terminal refers to the device installed on the side of the road or on the road section. Update the vehicle-side historical map based on at least one of the following: road obstacle information, road addition / deletion information, and operation boundary addition / deletion information.

[0005] In this embodiment, road obstacle information includes: pothole areas and pothole risk levels; determining road obstacle information for the traveled road segment based on road condition point cloud includes: identifying pothole areas in the road condition point cloud based on a visual recognition model, and determining the pothole size information and the distance between the pothole area and the road centerline; determining the pothole risk level based on the pothole size information and the distance; updating the vehicle-side historical map when any one or more of the road obstacle information, road addition / deletion information, and work boundary addition / deletion information are obtained includes: updating the pothole areas and marking the pothole risk level in the traveled road segment on the vehicle-side historical map when the pothole areas and pothole risk levels are obtained.

[0006] In this embodiment of the application, the pothole size information includes: pothole opening area, pothole maximum depth, and pothole average slope; determining the pothole risk level based on the pothole size information and interval distance includes: determining the pothole risk value based on the pothole opening area, pothole maximum depth, pothole average slope, and interval distance; and assessing the pothole risk level of the pothole area on the road based on the relationship between the pothole risk value and a preset risk threshold.

[0007] In this embodiment of the application, the pitting risk value is determined based on formula (1): (1) in, Indicates the risk value of pitting. Indicates the area of ​​the pit opening The risk value contribution function, Indicates the maximum depth of the pit The risk value contribution function, Indicates the average slope of the pit The risk value contribution function, Indicates interval distance The risk value contribution function, , , and These represent the preset weights of each contribution function.

[0008] In this embodiment, the preset risk threshold includes a first preset risk threshold and a second preset risk threshold greater than the first preset risk threshold. Assessing the pothole risk level of a road pothole area based on the relationship between the pothole risk value and the preset risk threshold includes: determining the pothole risk level of the road pothole area as low risk when the pothole risk value is less than the first preset risk threshold; determining the pothole risk level of the road pothole area as medium risk when the pothole risk value is greater than or equal to the first preset risk threshold and less than the second preset risk threshold, wherein the medium risk level is used to instruct vehicles to slow down; and determining the pothole risk level of the road pothole area as high risk when the pothole risk value is greater than the second preset risk threshold, wherein the high risk level is used to instruct vehicles to avoid the pothole.

[0009] The second aspect of this application provides a map updating method applied in the cloud. The updating method includes: acquiring road condition point clouds collected by a vehicle on a passing road segment, and road obstacle information determined by the vehicle based on the road condition point clouds of the passing road segment; processing the road condition point clouds to obtain road surface information of the passing road segment, and determining road surface addition / deletion information of the passing road segment based on the differential fusion result of the road surface information and the historical map in the cloud; acquiring operation boundary addition / deletion information, wherein the operation boundary addition / deletion information is determined by a roadside edge device based on the differential fusion result of the operation boundary update information and the historical map of the roadside edge device, and the operation boundary update information is determined by the roadside edge device processing vehicle trajectory data and environmental point clouds collected by the roadside edge device, wherein the roadside edge device refers to a device installed on the roadside or on the passing road segment; and updating the historical map in the cloud based on at least one of the road obstacle information, road surface addition / deletion information, and operation boundary addition / deletion information.

[0010] In this embodiment, processing the road condition point cloud to obtain road surface information of the traversed road segment, and determining the road surface addition / deletion information of the traversed road segment based on the differential fusion result of the road surface information and the historical map in the cloud includes: determining the preliminary road plane in the road condition point cloud based on the graph cut RANSAC algorithm; adjusting the attribution relationship between each data point in the road condition point cloud and the preliminary road plane, and determining the energy function value of each attribution relationship; selecting the attribution relationship with the smallest energy function value as the final attribution relationship; determining the in-plane data points belonging to the preliminary road plane among each data point according to the final attribution relationship, thereby obtaining the road surface information of the traversed road segment; and determining the road surface addition / deletion information of the traversed road segment based on the differential fusion result of the road surface information and the historical map in the cloud.

[0011] In this embodiment of the application, determining the energy function value of each attribution relationship includes: when the data point belongs to the preliminary road plane, determining the terrain point corresponding to the data point in the preset mine terrain model according to the spatial location of the data point; determining the elevation difference between the terrain point and the data point; and determining the energy function value of the attribution relationship according to the elevation difference of each data point.

[0012] In this embodiment, the energy function value is determined based on formula (2): (2) in, Represents the energy function value. This represents the set of data points in the road condition point cloud. Indicates the first The attribution relationship of each data point Indicates based on the first The data items are determined by the distance between each data point and the preliminary road plane. Indicates the first The data point and the The smoothing term is generated by the different attribution relationships of the data points. Indicates the first part belonging to the preliminary road plane Each data point is based on prior terms determined by elevation difference. , These represent the weight coefficients for each item.

[0013] A third aspect of this application provides a map updating method applied to a roadside edge, whereby the roadside edge refers to a device installed on the roadside or on the road segment traversed by a vehicle. The updating method includes: collecting environmental point clouds, acquiring vehicle trajectory data collected by the vehicle, and road obstacle information from the vehicle, wherein the road obstacle information is identified by the vehicle based on a visual recognition model in the collected road condition point cloud; processing the vehicle trajectory data and environmental point clouds to obtain work boundary update information for frequently traversed road segments, and determining work boundary addition / deletion information for frequently traversed road segments based on the work boundary update information and the differential fusion result of the roadside edge historical map; processing the vehicle's road condition point cloud in the cloud to obtain road surface information for the traversed road segment, and determining road surface addition / deletion information for the traversed road segment based on the road surface information and the differential fusion result of the cloud historical map; and updating the roadside edge historical map of the roadside edge based on at least one of the road obstacle information, road surface addition / deletion information, and work boundary addition / deletion information.

[0014] In this embodiment, processing vehicle trajectory data and environmental point clouds to obtain work boundary update information for frequently traveled road sections, and determining the addition or deletion information of work boundaries for frequently traveled road sections based on the work boundary update information and the differential fusion result of the historical map at the roadside edge, includes: processing vehicle trajectory data based on a kernel density estimation algorithm to obtain a heat map of the spatial distribution of vehicle trajectories; delineating geofences based on vehicle trajectories with heat values ​​higher than a threshold in the heat map, wherein the road sections delineated by the geofences are frequently traveled road sections; determining the point clouds to be processed that fall into the geofences based on the spatial location of the environmental point clouds; separating the road surface point clouds in the point clouds to be processed based on the ray-mapping method; generating the work boundary of the road surface point clouds based on the Alpha Shape algorithm as the work boundary update information; and determining the addition or deletion information of work boundaries for frequently traveled road sections based on the work boundary update information and the differential fusion result of the historical map at the roadside edge.

[0015] In this embodiment of the application, the vehicle end is a mining vehicle, the vehicle end historical map is a mining historical map saved by the mining vehicle, the cloud historical map is a mining historical map saved in the cloud, and the roadside edge end historical map is a mining historical map saved at the roadside edge.

[0016] A fourth aspect of this application provides a map update system, comprising: a vehicle-mounted terminal for collecting vehicle trajectory data and road condition point clouds of the route traveled by the vehicle; the vehicle-mounted terminal is further configured to determine road obstacle information of the route traveled based on the road condition point clouds, and send the road obstacle information to a roadside edge terminal and / or a cloud terminal; the cloud terminal is configured to acquire the road condition point clouds collected by the vehicle-mounted terminal on the route traveled, process the road condition point clouds to obtain road surface information of the route traveled, and determine road surface addition / deletion information of the route traveled based on the differential fusion result of the road surface information and the historical map in the cloud terminal; the cloud terminal is further configured to send the road surface addition / deletion information to the vehicle-mounted terminal and / or the roadside edge terminal; the roadside edge terminal is configured to collect environmental point clouds. The system acquires vehicle trajectory data collected by the vehicle and processes the vehicle trajectory data and environmental point cloud to obtain the work boundary update information of the vehicle in frequently traveled road sections. Based on the differential fusion results of the work boundary update information and the historical map of the roadside edge terminal, the system determines the work boundary addition and deletion information of the frequently traveled road sections. The roadside edge terminal refers to the device installed on the roadside or on the road section. The roadside edge terminal is also used to send the work boundary addition and deletion information to the vehicle and / or the cloud. The vehicle, the roadside edge terminal, and the cloud can update their respective historical maps based on at least one of the road obstacle information, road addition and deletion information, and work boundary addition and deletion information.

[0017] The fifth aspect of this application provides a map updating apparatus, including a processor configured to perform the map updating method described above.

[0018] A sixth aspect of this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the map update method described above.

[0019] Through the above technical solution, the vehicle can determine road obstacle information based on road condition point clouds, thereby instantly identifying road obstacles and enabling the vehicle to perform real-time avoidance, deceleration, and other operations based on road obstacles. Since the road itself has a long change cycle and the requirement for timely updates is not high, the road condition point clouds collected by the vehicle are processed in the cloud to determine road information. Road addition / deletion information is obtained through differential fusion of the road information and historical maps in the cloud for map updates. Determining road addition / deletion information in the cloud reduces the computational load on the vehicle and allows the cloud to integrate road condition point clouds from different vehicles on different road sections to determine road addition / deletion information. The roadside edge device is installed on the roadside or on the road and determines the operation boundary update information based on the vehicle trajectory data and the environmental point cloud collected by the roadside edge device. Then, based on the operation boundary update information and the road... The differential fusion results of the historical map at the side edge are used to determine the operation boundary. The operation boundary can reflect the production status and engineering layout of the mine operation. The frequency of operation boundary changes and adjustments is generally higher than that of road changes, and it does not require the timeliness of road obstacle detection. Therefore, the operation boundary addition and deletion information can be accurately determined with low latency through the roadside edge. The vehicle itself can determine road obstacle information and receive road addition and deletion information and operation boundary addition and deletion information, thereby updating the vehicle's historical map. Due to the poor communication conditions in the mining environment, in order to achieve efficient bandwidth utilization and meet real-time communication, the above-mentioned road obstacle information, road addition and deletion information, and operation boundary addition and deletion information are all determined based on the differential fusion results. Therefore, only the information added or subtracted relative to the historical map is included, which greatly reduces the amount of data transmission and storage pressure, and ensures the efficiency, consistency and security of map version management.

[0020] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0021] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 The illustration shows a schematic flowchart of a map updating method according to an embodiment of this application; Figure 2 This illustration schematically shows a flowchart of a method for determining road obstacle information according to an embodiment of this application; Figure 3The schematic diagram illustrates a flowchart of another map updating method provided according to an embodiment of this application; Figure 4 The illustration shows a flowchart of a method for determining road surface addition / deletion information according to an embodiment of this application; Figure 5 The schematic diagram illustrates a flowchart of another map updating method provided according to an embodiment of this application; Figure 6 The illustration shows a flowchart of a method for determining job boundary addition / deletion information according to an embodiment of this application; Figure 7 The diagram illustrates the internal structure of a computer device according to an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0023] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0024] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0025] The acquisition, transmission, storage, use, and processing of data in this application comply with relevant laws and regulations. Furthermore, it should be noted that certain software, components, models, and other existing industry solutions may be mentioned in the embodiments of this application. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0026] Current autonomous driving maps for mines primarily rely on offline static maps, typically using periodic data collection and manual annotation to identify changing areas. However, such methods struggle to adapt to the rapid dynamic changes in the mining environment. To achieve high-precision mine map creation, 3D reconstruction can be performed using laser point clouds collected in real-time by vehicles. However, this method often faces high computational loads, and due to poor communication conditions in mining areas, even cloud computing can lead to untimely map updates due to communication delays or short-term failures. Based on the above analysis, this application proposes a map updating method that collaboratively updates maps through vehicle-side, cloud-side, and roadside edge devices. These devices are responsible for determining different map update information to achieve timely and accurate map updates.

[0027] Figure 1 A schematic flowchart of a map update method according to an embodiment of this application is shown. Figure 1 As shown, in one embodiment of this application, a map updating method is provided, applied to a vehicle. The vehicle refers to the computing unit, sensors (LiDAR, perception camera, millimeter-wave radar, etc.), actuators, and vehicle communication module included in the vehicle. The vehicle is responsible for environmental perception, local calculation, decision control, and communication with roadside edge devices and the cloud. The map updating method includes the following steps: S102. Collect vehicle trajectory data and road condition point cloud data of the vehicle on the route; S104. Determine road obstacle information for the route based on road condition point cloud; S106. Obtain road surface addition and deletion information for the route. The road surface addition and deletion information for the route is determined by the cloud based on the differential fusion result of the road surface information corresponding to the road condition point cloud and the historical map in the cloud.

[0028] The cloud can be a remote cloud computing center with powerful global computing and storage capabilities, enabling the aggregation, in-depth analysis, model training, global scheduling and management of massive amounts of data.

[0029] S108. Obtain information on the addition or deletion of the operation boundary. The information on the addition or deletion of the operation boundary is determined by the roadside edge terminal based on the differential fusion result of the operation boundary update information and the historical map of the roadside edge terminal. The information on the update of the operation boundary is determined by the roadside edge terminal processing vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. The roadside edge terminal refers to the device installed on the side of the road or on the road section.

[0030] Roadside edge devices can be edge servers deployed on the roadside (such as streetlights and traffic light poles), close to vehicles, and can provide localized, near real-time computing, storage, and communication services.

[0031] S110. Update the vehicle-side historical map based on at least one of the following: road obstacle information, road addition / deletion information, and operation boundary addition / deletion information.

[0032] In the map update method provided in this application embodiment, the vehicle determines road obstacle information of the traversed road segment based on road condition point cloud, thereby enabling real-time identification of road obstacles and allowing the vehicle to perform real-time avoidance, deceleration, and other operations based on road obstacles. Since the road itself has a long change cycle and the requirement for timely updates is not high, the road condition point cloud collected by the vehicle is processed in the cloud to determine road information. Road addition / deletion information is obtained through differential fusion of the road information and the historical map in the cloud for map updates. Determining road addition / deletion information in the cloud reduces the computational load on the vehicle and allows the cloud to integrate road condition point clouds from different vehicles traversing different road segments to determine this information. The roadside edge terminal is installed on the roadside or on the road and determines the operation boundary update information based on the vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. This is then determined based on the differential fusion results of the operation boundary update information and the historical map of the roadside edge terminal. The work boundary reflects the production status and engineering layout of mining operations. The frequency of changes to the work boundary is generally higher than that of road changes, and it does not require the timeliness of road obstacle detection. Therefore, the work boundary addition / deletion information is accurately determined with low latency through the roadside edge. The vehicle itself can determine road obstacle information and receive road and work boundary addition / deletion information, thereby updating the vehicle's historical map. Due to the poor communication conditions in the mining environment, to achieve efficient bandwidth utilization and meet real-time communication requirements, the aforementioned road obstacle information, road addition / deletion information, and work boundary addition / deletion information are all determined based on differential fusion results. Therefore, only information added or subtracted relative to the historical map is included, significantly reducing data transmission volume and storage pressure, while ensuring the efficiency, consistency, and security of map version management. In summary, the map update method provided in this application can reduce data transmission and storage pressure, and rationally allocate computing resources to achieve timely and accurate map updates.

[0033] Specifically, the road obstacle information on the vehicle can be updated frequently, in units of seconds or minutes. The road obstacle information can include road obstacle elements such as fallen rocks / minerals, obstacles, and potholes.

[0034] In some embodiments of this application, road obstacle information includes: pothole areas and pothole risk levels. See also Figure 2 Based on the road condition point cloud, the road obstacle information for the traversed road segment can include: S202. Identify road pothole areas in the road condition point cloud based on the visual recognition model, and determine the pothole size information and the distance between the road pothole area and the road centerline. The risk level of a pit is determined based on its size and spacing. When one or more of the following are obtained: road obstacle information, road addition / deletion information, and work boundary addition / deletion information, updating the vehicle-side historical map includes: Once the pothole areas and their risk levels are known, the pothole areas are updated on the historical map of the vehicle, and the risk levels of the pothole areas are marked.

[0035] The above embodiments determine pothole areas on the road based on a visual recognition model, and assess the pothole risk level based on the pothole size information and the distance between the pothole area and the road centerline. This allows the vehicle to update pothole areas and mark their risk levels on the historical map of the route traveled, enabling the vehicle to take appropriate obstacle avoidance or warning actions during the route traveled. Specifically, the visual recognition model can be a PointNet++, DGCNN, Point Transformer, or similar model.

[0036] See Figure 2 In some embodiments of this application, the pit size information includes: pit opening area, maximum pit depth, and average pit slope; determining the pit risk level based on the pit size information and interval distance includes: S204. Determine the pit risk value based on the pit opening area, pit maximum depth, pit average slope, and interval distance. S206. Assess the pothole risk level of the pothole area on the road based on the relationship between the pothole risk value and the preset risk threshold.

[0037] Based on the above steps, the pothole risk value reflects the degree to which a pothole obstructs vehicle passage through its opening area; its maximum depth reflects the risk of bottoming out or severe jolting when the vehicle passes through it; and its average slope reflects the steepness of the pothole terrain. A steeper slope results in more drastic changes in vehicle posture when entering or exiting the pothole, increasing the likelihood of cargo overturning or suspension system damage. The average slope and spacing of the potholes also reflect the probability that they lie on the vehicle's normal driving trajectory. Therefore, the pothole risk value comprehensively assesses the risk posed by potholes to vehicle passage by considering all these factors.

[0038] In some embodiments of this application, the pitting risk value is determined based on formula (1): (1) in, Indicates the risk value of pitting. Indicates the area of ​​the pit opening The risk value contribution function, Indicates the maximum depth of the pit The risk value contribution function, Indicates the average slope of the pit The risk value contribution function, Indicates interval distance The risk value contribution function, , , and These represent the preset weights of each contribution function.

[0039] Specifically, opening area The risk value contribution function can be, for example, as follows: ,in, This indicates the preset maximum effective area threshold. Maximum pit depth. The risk value contribution function can be, for example, as follows: ,in, Indicates the depth threshold. Average slope of the pit. The risk value contribution function can be, for example, as follows: ,in, Indicates the slope threshold. Interval distance. The risk value contribution function can be, for example, as follows: ,in, These are preset constants. Based on the aforementioned dent size information and interval distance, the contribution function can achieve a normalized representation of the dent risk value. The preset weights of each contribution function can be obtained through regression analysis using historical accident data or expert driving experience, and the preset weights of each contribution function can satisfy the constraint that the summation equals 1. As an example, , , and .

[0040] In some embodiments of this application, the preset risk thresholds include: a first preset risk threshold and a second preset risk threshold greater than the first preset risk threshold; assessing the pothole risk level of a road pothole area based on the relationship between the pothole risk value and the preset risk thresholds includes: determining the pothole risk level of the road pothole area to be low-risk when the pothole risk value is less than the first preset risk threshold; determining the pothole risk level of the road pothole area to be medium-risk when the pothole risk value is greater than or equal to the first preset risk threshold and less than the second preset risk threshold, wherein the medium-risk level is used to inform the vehicle to slow down; and determining the pothole risk level of the road pothole area to be high-risk when the pothole risk value is greater than the second preset risk threshold, wherein the high-risk level is used to inform the vehicle to avoid the pothole. Based on the above steps, pothole risk level assessment based on pothole risk value is achieved. The pothole risk level can be updated in the vehicle's historical map to inform the vehicle to take corresponding actions to avoid the risk, significantly improving the safety and reliability of autonomous driving in mines.

[0041] As an example, the risk value of the pit is in the range of 0 to 1, the first preset risk threshold can be, for example, 0.3, and the second risk threshold can be, for example, 0.6.

[0042] Low risk level (0 ≤ <0.3): Vehicles can pass normally; only recording is required. Medium risk level (0.3 ≤ <0.6): Vehicles need to slow down before passing and report the incident to the cloud. High risk level ( ≥ 0.6): Vehicles should avoid the area, and a warning should be broadcast immediately.

[0043] This model enables the system to quantify the risk of dynamically detected potholes in real time, thereby triggering different levels of warnings and response strategies. Understandably, vehicle-mounted broadcast warnings can include broadcasting the pothole area and its risk level to other vehicles, roadside edges, and the cloud.

[0044] In some embodiments of this application, the vehicle-side can also achieve cross-modal fusion detection of vision and radar through cameras and radar. The vehicle-side can form a 360-degree fusion perception network centered on the vehicle through spatial calibration and temporal synchronization with lidar and surround-view cameras, continuously monitoring its coverage area. These sensors can collect environmental information around the vehicle in real time. Cameras are responsible for capturing image information from the front, rear, and sides of the vehicle; radar and millimeter-wave radar data provide distance and speed information; lidar collects high-precision, high-resolution 3D point cloud data of the surrounding environment; the vehicle-side performs noise reduction (e.g., removing radar clutter), formatting (e.g., unifying timestamps and coordinate systems), and compression on the collected raw data, extracting key structured information (e.g., the position, speed, and type of targets), and standardizing the data collected by different sensors. Using the preprocessed data, data from different sensors (e.g., radar point clouds and camera images) are fused through data fusion algorithms (e.g., Kalman filtering) to generate a more comprehensive and accurate environmental description for subsequent analysis and processing. Simultaneously, the preprocessing of vehicle-side data greatly improves data quality and significantly reduces the amount of data and bandwidth required for transmission to the cloud.

[0045] In some embodiments of this application, the vehicle can acquire images of the road segment it is traveling through via a camera and acquire road condition point clouds of the road segment via radar; the road segment images and road condition point clouds are fused according to the spatiotemporal calibration relationship of the camera and radar to obtain a three-dimensional road condition model; the preliminary position and type of obstacle are determined in the three-dimensional road condition model using image recognition models such as YOLOv5s; the "view cone" region where the obstacle is located is determined in the three-dimensional road condition model based on the preliminary position of the obstacle, and the Alpha Shape algorithm is applied to extract the boundary of the point cloud within this "view cone" region, thereby obtaining a rough three-dimensional outline and spatial position of the obstacle that can be used as road obstacle information for the vehicle.

[0046] See Figure 3 This application also provides a map update method, applied in the cloud, which includes: S302. Obtain road condition point cloud data collected by the vehicle on the route it is traveling; S304. Obtain road obstacle information determined by the vehicle based on the road condition point cloud of the route traveled; S306. Process the road condition point cloud to obtain the road surface information of the traveled road segment, and determine the road surface addition or deletion information of the traveled road segment based on the differential fusion results of the road surface information and the historical map in the cloud. S308. Obtain information on the addition or deletion of the work boundary. The information on the addition or deletion of the work boundary is determined by the roadside edge terminal based on the differential fusion result of the work boundary update information and the historical map of the roadside edge terminal. The work boundary update information is determined by the roadside edge terminal processing vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. The roadside edge terminal refers to the device installed on the side of the road or on the road section. S310. Update the cloud-based historical map based on at least one of the following: road obstacle information, road addition / deletion information, and work boundary addition / deletion information.

[0047] In the map update method provided in this application embodiment, given that roads themselves, as geographical elements, have a long change cycle and low requirements for real-time updates, road surface information is determined by processing road condition point clouds collected by vehicles in the cloud. Road surface addition / deletion information is obtained through differential fusion of the road surface information and historical maps in the cloud for map updates. Determining road surface addition / deletion information in the cloud reduces the computational load on the vehicle and allows the cloud to integrate road condition point clouds from different vehicles traveling on different road segments. The vehicle determines road obstacle information based on the road condition point clouds, enabling real-time obstacle identification and allowing the vehicle to perform real-time avoidance, deceleration, and other operations. The roadside edge device is installed on the roadside or on the road and determines the operation boundary update information based on vehicle trajectory data and environmental point clouds collected by the roadside edge device. This work boundary update information is then determined based on the differential fusion results of the operation boundary update information and historical maps from the roadside edge device. The work boundary reflects the production status and engineering layout of mining operations. The frequency of changes to the work boundary is generally higher than that of road changes, and it does not require the timeliness of road obstacle detection. Therefore, the addition or deletion of work boundary information is accurately determined with low latency through the roadside edge. The cloud can integrate road obstacle information, road addition / deletion information, and work boundary addition / deletion information to update the cloud historical map. Due to the poor communication conditions in the mining environment, to achieve efficient bandwidth utilization and meet real-time communication requirements, the aforementioned road obstacle information, road addition / deletion information, and work boundary addition / deletion information are all determined based on differential fusion results. Therefore, only information added or subtracted relative to the historical map is included, which significantly reduces data transmission volume and storage pressure, and ensures the efficiency, consistency, and security of map version management.

[0048] In some embodiments of this application, the road condition point cloud collected by the vehicle on the route is a point cloud of the environment in which the vehicle is located. Therefore, the road condition point cloud may include the point cloud of the route and the point cloud of the road environment. The cloud can also identify static targets such as loading areas, unloading areas and fixed facilities in the mine in the road condition point cloud through visual recognition models such as PointNet++; determine the static target addition and deletion information of the route by the differential fusion result of the identified static targets and the historical map in the cloud; step S310 may include: updating the historical map in the cloud based on at least one of road obstacle information, static target addition and deletion information, road surface addition and deletion information and work boundary addition and deletion information. The static target addition and deletion information may also be sent to the vehicle to update the historical map of the vehicle and / or sent to the roadside edge to update the historical map of the roadside edge.

[0049] See Figure 4 In some embodiments of this application, step S306 may include: S402. Determine the preliminary road plane in the road condition point cloud based on the graph cut RANSAC algorithm.

[0050] Specifically, the graph cut RANSAC algorithm initially estimates the parameters of the planar model using the minimum sample set of the road condition point cloud. The minimum sample set may include three data points in the road condition point cloud that constitute a plane.

[0051] S404. Adjust the attribution relationship between each data point in the road condition point cloud and the preliminary road plane, and determine the energy function value of each attribution relationship; S406. Select the attribution relationship with the smallest energy function value as the final attribution relationship; S408. Based on the final attribution relationship, determine the in-plane data points belonging to the preliminary road plane among each data point to obtain the road surface information of the traversed road segment; S410. Determine the road surface addition and deletion information for the route based on the differential fusion results of road surface information and cloud-based historical maps.

[0052] To address the characteristics of high noise and high feature coefficients in mine point clouds, this application employs a graph-cut RANSAC algorithm optimized for mining scenarios to determine the initial road plane. Based on minimizing the energy function value, it identifies data points in the road condition point cloud belonging to the initial road plane, thereby determining the road surface information of the traversed road segment. Furthermore, based on the differential fusion results of the road surface information and the historical cloud map, it determines the addition or deletion of road surface information for the traversed road segment. Understandably, the road surface information of the traversed road segment includes the road surface formed by in-plane data points belonging to the initial road plane.

[0053] In some embodiments of this application, step S404, determining the energy function value for each attribution relationship, may include: when a data point belongs to the preliminary road plane, determining the corresponding terrain point in a preset mine terrain model based on the spatial location of the data point; determining the elevation difference between the terrain point and the data point; and determining the energy function value for the attribution relationship based on the elevation difference of each data point. The energy function value determined based on the elevation difference of the data points ensures that the in-plane data points belonging to the preliminary road plane are spatially continuous and conform to the interior point set of the plane model, effectively combating noise interference.

[0054] In some embodiments of this application, the energy function value is determined based on formula (2): (2) in, Represents the energy function value. This represents the set of data points in the road condition point cloud. Indicates the first The attribution relationship of each data point Indicates based on the first The data items are determined by the distance between each data point and the preliminary road plane. Indicates the first The data point and the The smoothing term is generated by the different attribution relationships of the data points. Indicates the first part belonging to the preliminary road plane Each data point is based on prior terms determined by elevation difference. , These represent the weight coefficients for each item.

[0055] Specifically, Taking different values ​​can represent the first Different attribution relationships of data points, for example Indicates the first The attribution relationship of the data point is the first... The data point belongs to the preliminary road plane, that is, the first data point. Each data point is an interior point. Indicates the first The attribution relationship of the data point is the first... The data point does not belong to the preliminary road plane, that is, the first data point. Each data point is an outlier. Data item Measurable The degree of fit between each data point and the initial road plane. If the data points are marked as inliers ( However, a large distance between the data points and the initial road surface will result in high costs; conversely, a small distance will result in low costs. (Smoothing term) Encourage spatially adjacent data points (e.g., the first...) The data point and the (Number of data points) have the same label (both are interior points or both are exterior points). If two adjacent points have different affiliations, a penalty cost will be incurred on the edge connecting them. Prior terms For example, ,in, For preset coefficients, Indicates the first The elevation difference of each data point. Based on prior terms. This can reduce the likelihood that data points with large elevation differences will be identified as interior points.

[0056] See Figure 5 This application also provides a map updating method applied to the roadside edge, where the roadside edge refers to a device installed on the roadside or on the road segment through which the vehicle travels. The updating method includes: S502. Collect environmental point cloud data and obtain vehicle trajectory data collected by the vehicle terminal.

[0057] Specifically, the vehicle-side trajectory data can be determined based on satellite positioning signals obtained from the satellite positioning and navigation device and vehicle-side motion data collected by inertial sensors. The vehicle-side trajectory data may also include vehicle outline dimension information.

[0058] S504. Obtain road obstacle information from the vehicle, wherein the road obstacle information is identified by the vehicle based on a visual recognition model in the collected road condition point cloud. S506. Process vehicle trajectory data and environmental point cloud to obtain the work boundary update information of the vehicle in frequently traveled road sections, and determine the addition or deletion information of the work boundary of frequently traveled road sections based on the differential fusion results of the work boundary update information and the historical map at the roadside edge. S508: Obtain road surface information of the traveled road segment by processing the road condition point cloud on the vehicle in the cloud, and determine the addition or deletion of road surface information of the traveled road segment based on the differential fusion result of the road surface information and the historical map in the cloud. S510. Update the historical map of the roadside edge based on at least one of the following: road obstacle information, road addition / deletion information, and work boundary addition / deletion information.

[0059] In the map update method provided in this application embodiment, the roadside edge terminal is installed on the roadside or on the road, and determines the work boundary update information based on vehicle trajectory data and environmental point clouds collected by the roadside edge terminal. Then, it is determined based on the differential fusion results of the work boundary update information and the historical map of the roadside edge terminal. The work boundary can reflect the production status and engineering layout of mining operations. The frequency of work boundary changes and adjustments is generally higher than the frequency of road changes, and it does not require the timeliness of road obstacle detection. Therefore, the roadside edge terminal is used to accurately determine the addition and deletion information of the work boundary with low latency. Given the long change cycle of the road itself and the low requirement for real-time updates, the road surface information is determined by processing the road condition point clouds collected by the vehicle terminal in the cloud. The road surface information is then differentially fused with the historical map in the cloud to obtain road surface addition and deletion information for map updates. Determining road surface addition and deletion information in the cloud reduces the computational load on the vehicle terminal, and the cloud can also integrate road condition point clouds from different vehicles traveling on different road sections to determine the road surface addition and deletion information. The cloud can integrate road obstacle information, road surface addition and deletion information, and work boundary addition and deletion information to update the historical map in the cloud. The vehicle-mounted system determines road obstacle information based on road condition point clouds, enabling real-time obstacle identification and actions such as obstacle avoidance and deceleration. Due to the poor communication conditions in mining environments, to achieve efficient bandwidth utilization and meet real-time communication requirements, the aforementioned road obstacle information, road surface addition / deletion information, and work boundary addition / deletion information are all determined based on differential fusion results. Therefore, only information added or removed relative to historical maps is included, significantly reducing data transmission volume and storage pressure, while ensuring the efficiency, consistency, and security of map version management.

[0060] Specifically, the information on the addition and deletion of work boundaries at the roadside edge can be updated periodically on a daily or weekly basis. This information may include additions or deletions of work area boundaries, step lines, and retaining walls. The roadside edge refers to a heterogeneous resource network composed of edge nodes in the vehicle-side-cloud collaborative update technology. It is crucial for real-time information interaction in this technology and can achieve data transmission between multiple devices based on various wireless communication methods such as 5G and Wi-Fi. The roadside edge can issue transportation instructions from the cloud, receive data from vehicles, and utilize its computing and storage resources to quickly detect changes in map features. It then synchronizes the updated information to other roadside edge nodes, vehicles, and the cloud, completing map updates quickly and accurately with low latency, thus ensuring the efficiency of map updates.

[0061] See Figure 6 In some embodiments of this application, step S506 may include: S602. Process vehicle trajectory data based on kernel density estimation algorithm to obtain a heat map of the spatial distribution of vehicle trajectory.

[0062] Specifically, the vehicle trajectory data can be, for example, preprocessed trajectory data recorded by the vehicle in the past hour.

[0063] S604. Geofencing is defined based on the vehicle trajectories with heat values ​​higher than the threshold in the heat map, wherein the geofenced sections are frequently traveled sections.

[0064] Specifically, a geofence can be represented by a minimum bounding rectangle or a convex hull.

[0065] S606. Determine the point clouds to be processed that fall into the geofence based on the spatial location of the environmental point cloud. S608. Separation of road surface point cloud from point cloud to be processed based on ray tracing method; S610. The operation boundary of the road surface point cloud generated based on the Alpha Shape algorithm is used as the operation boundary update information. S612. Determine the addition or deletion information of the operation boundary for frequently traversed road sections based on the differential fusion results of the operation boundary update information and the historical map at the roadside edge.

[0066] Based on the above steps, the geofence defined by the processor vehicle trajectory can reflect the frequently traveled road segments. Then, the point cloud to be processed within the geofence area can be centrally processed based on the ray method and Alpha Shape algorithm to determine the operation boundary update information. This effectively obtains the operation boundary addition and deletion information while reducing the computational load at the roadside edge.

[0067] In some embodiments of this application, step S602 may include: performing outlier removal and smoothing filtering on the vehicle trajectory data based on Mahalanobis distance to obtain preprocessed trajectory data; and processing the preprocessed trajectory data based on a kernel density estimation algorithm to obtain a heat map of the spatial distribution of the vehicle trajectory.

[0068] In some embodiments of this application, step S608 may include voxel downsampling of the point cloud to be processed to obtain a sampled point cloud; thereby reducing the amount of data to be processed, and then separating ground feature clusters (such as blast piles and equipment) in the sampled point cloud based on the DBSCAN algorithm to obtain a preliminary ground point cloud; and separating road surface point clouds in the preliminary ground point cloud based on the ray method.

[0069] In some embodiments of this application, "vehicle-side" refers to a mining vehicle, "vehicle-side historical map" is a mining historical map saved by the mining vehicle, "cloud-based historical map" is a mining historical map saved in the cloud, and "roadside edge-side historical map" is a mining historical map saved at the roadside edge.

[0070] It should be understood that although the steps in the flowcharts provided in this application are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0071] This application also provides a map update system, including: a vehicle-mounted terminal, a cloud-based terminal, and a roadside edge terminal. The vehicle-mounted terminal is used to collect vehicle trajectory data and road condition point cloud data of the road segment it is traveling on; the vehicle-mounted terminal is also used to determine road obstacle information of the road segment it is traveling on based on the road condition point cloud data, and send the road obstacle information to the roadside edge terminal and / or the cloud-based terminal. The cloud is used to acquire road condition point clouds collected by the vehicle on the route, process the road condition point clouds to obtain road surface information of the route, and determine the road surface addition or deletion information of the route based on the differential fusion results of the road surface information and the historical map in the cloud; the cloud is also used to send the road surface addition or deletion information to the vehicle and / or the roadside edge end. The roadside edge terminal is used to collect environmental point cloud data and vehicle trajectory data collected by the vehicle terminal. It is also used to process the vehicle trajectory data and environmental point cloud data to obtain the operation boundary update information of the vehicle in frequently traveled road sections. Based on the differential fusion results of the operation boundary update information and the historical map of the roadside edge terminal, the addition and deletion information of the operation boundary of the frequently traveled road sections is determined. The roadside edge terminal refers to the device installed on the roadside or on the road section. The roadside edge terminal is also used to send the operation boundary addition and deletion information to the vehicle terminal and / or the cloud. The vehicle-side, roadside edge, and cloud-based systems update their respective historical maps based on at least one of the following: road obstacle information, road addition / deletion information, and work boundary addition / deletion information.

[0072] In some embodiments of this application, road obstacle information, road addition / deletion information, and work boundary addition / deletion information can be, for example, incremental update packets in JSON format. As an example, the following exemplifies the process of map updates implemented through three-terminal interaction in a map update system: After a vehicle detects an obstacle, the generated incremental packet is directly broadcast to surrounding vehicles and roadside edge terminals via 5G-V2X. After initial aggregation and verification by the roadside edge terminals, the aggregated packet is uploaded to the cloud. Update packets generated primarily by the roadside edge terminals are uploaded to the cloud periodically or triggered. The cloud acts as a global data fusion center, receiving and processing all uplink incremental packets such as road obstacle information, road addition / deletion information, and work boundary addition / deletion information; it maintains a four-dimensional spatiotemporal cube (i.e., three-dimensional space and one-dimensional time) database to manage all historical versions of map elements, ensuring that updates do not cause spatiotemporal logic conflicts. The cloud can obtain the latest global map incremental packet based on the above-mentioned uplink incremental packets and distribute it as needed according to the location and task requirements of the vehicle and roadside edge terminals. For example, prioritizing the distribution of update packages for a specific transport route and its surrounding areas to vehicles operating in that area can significantly save communication bandwidth.

[0073] In some embodiments of this application, to ensure update quality, the map update methods executed by the cloud, roadside edge, and vehicle may also include automated map verification mechanisms.

[0074] Static layer verification is performed in the cloud, including geometric accuracy verification: an improved 95th quantile Hausdorff distance is used to quantify the differences between road surface additions / removals and the new and old versions of road boundaries in the cloud historical map. This method is insensitive to outliers and more robustly assesses the similarity of the overall shape, ensuring that updates do not introduce significant geometric biases.

[0075] Static layer verification also includes topology logic verification: a depth-first search (DFS) algorithm is used to check road network connectivity to ensure that no isolated road segments are created due to updates; the SWEEP algorithm is used to check for path topology errors. Path topology errors may include isolated paths that do not connect with other paths or work areas.

[0076] Quasi-dynamic layer verification is performed at the roadside edge, focusing on verifying the rationality of changes in the work boundary. Based on the updated work boundary information and the historical work boundaries in the historical map at the roadside edge, the overlap area ratio (IoU) of the two work boundaries is calculated, and its expansion direction is verified to be consistent with the production plan.

[0077] Quasi-dynamic layer verification performed at the roadside edge can also include checking whether the curvature change of the road boundary in the operation boundary update information is smooth, preventing the curvature of the road boundary from exceeding the curvature threshold, avoiding unreasonable sharp corners, and ensuring that vehicles can drive smoothly.

[0078] Vehicle-side dynamic layer verification focuses on verifying the real-time performance of the perception algorithm. The vehicle-side verification primarily determines the recall, precision, and F1-score of obstacle recognition based on the visual recognition model within the road obstacle information determined by the road condition point cloud. It also rigorously measures the delay from collecting the road condition point cloud from the vehicle to updating the vehicle's historical map and issuing a warning to ensure timeliness.

[0079] When significant abnormal changes are detected at the vehicle end, roadside edge end, and cloud end, the system will automatically trigger a manual review process. Reviewers can use AR (augmented reality) devices to overlay and compare road obstacle information, road addition / deletion information, or work boundary addition / deletion information with the actual scene, completing the verification work intuitively and efficiently, forming a quality closed loop of "automation as the main method and manual assistance as the auxiliary method".

[0080] In summary, the map update method provided in this application achieves breakthrough improvements in multiple dimensions through hierarchical dynamic updates and a collaborative network architecture. In terms of efficiency, it achieves highly efficient hierarchical updates. The update cycle for the vehicle-side dynamic layer (road obstacle information) can reach the second level, while the update cycle for the roadside edge layer (work boundary addition / deletion information) can reach the hour or minute level. Through differential incremental update technology, only changed data (i.e., road obstacle information, road addition / deletion information, and work boundary addition / deletion information) is transmitted, reducing network transmission load by more than 70% and significantly improving update response speed. Regarding accuracy, through multi-source sensor fusion and improved algorithms (RANSAC and PointNet++ optimized for mining, and optimized work area boundary extraction strategies, etc.), the extraction accuracy of key elements such as road boundaries and work area boundaries is effectively improved. In terms of resource utilization, a "vehicle-edge-cloud" collaborative computing approach is adopted to rationally allocate computing tasks. The roadside edge layer handles quasi-dynamic layer processing, the vehicle-side layer is responsible for real-time dynamic layer perception, and the cloud focuses on global fusion. This architecture effectively reduces cloud load and vehicle-side computing pressure, optimizing resource utilization. In terms of reliability, a quality assurance system has been established that includes automated verification (Hausdorff distance geometry verification, topology logic checking) and human-machine collaborative review (AR-assisted verification). Risk assessment and graded early warning for dynamic obstacles significantly improve active safety. Practical applications show that this solution can greatly reduce emergency operations caused by map lag, improve mining area transportation efficiency and the overall reliability of the autonomous driving system, and provide solid technical support for unmanned mining operations.

[0081] This application also provides a map updating device, including a processor configured to execute the map updating method provided according to this application. Specifically, the map updating device can be installed at a vehicle, a cloud, or a roadside edge.

[0082] This application also provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform a map update method according to this application.

[0083] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 7 As shown. The computer device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The database stores one or more of the following: point cloud data, historical maps, road obstacle information, road surface addition / deletion information, and work boundary addition / deletion information. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements a map update method.

[0084] Those skilled in the art will understand that Figure 7 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0085] In one embodiment, the map updating apparatus provided in this application can be implemented as a computer program, which can be implemented in various ways, such as... Figure 7 The system operates on the computer device shown. The computer device's memory can store the various program modules that make up the map updating apparatus. The computer program, composed of the various program modules, causes the processor to execute the steps in the map updating method provided in the embodiments of this application described in this specification.

[0086] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0087] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0088] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0089] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0090] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0091] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0092] Computer-readable media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0093] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0094] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A map updating method, characterized in that, When applied to the vehicle side, the update method includes: Collect vehicle trajectory data and road condition point cloud data of the vehicle on the route it travels; The road obstacle information of the traversed road section is determined based on the road condition point cloud; Obtain road surface addition and deletion information for the traveled road segment. The road surface addition and deletion information for the traveled road segment is determined by the cloud based on the differential fusion result of the road surface information corresponding to the road condition point cloud and the historical map in the cloud. The information on the addition and deletion of the operation boundary is obtained. The information on the addition and deletion of the operation boundary is determined by the roadside edge terminal based on the differential fusion result of the operation boundary update information and the historical map of the roadside edge terminal. The information on the update of the operation boundary is determined by the roadside edge terminal processing the vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. The roadside edge terminal refers to the device installed on the roadside or on the road section. The vehicle-side historical map is updated based on at least one of the road obstacle information, the road addition / deletion information, and the operation boundary addition / deletion information.

2. The map updating method according to claim 1, characterized in that, The road obstacle information includes: pothole areas and pothole risk levels; The step of determining the road obstacle information of the traversed road segment based on the road condition point cloud includes: Based on a visual recognition model, pothole areas in the road condition point cloud are identified, and the pothole size information and the distance between the pothole area and the road centerline are determined. The risk level of the pit is determined based on the pit size information and the interval distance; When one or more of the road obstacle information, the road addition / deletion information, and the operation boundary addition / deletion information are obtained, updating the vehicle-side historical map includes: Given the pothole areas and pothole risk levels on the road, the pothole areas are updated in the historical map of the vehicle, and the pothole risk levels of the pothole areas are marked.

3. The map updating method according to claim 2, characterized in that, The pit size information includes: pit opening area, pit maximum depth, and pit average slope. The step of determining the pit risk level based on the pit size information and the interval distance includes: The risk value of the pit is determined based on the pit opening area, the maximum depth of the pit, the average slope of the pit, and the interval distance. The pothole risk level of the pothole area on the road is assessed based on the relationship between the pothole risk value and the preset risk threshold.

4. The map updating method according to claim 3, characterized in that, The risk value of pitting is determined based on formula (1): ； （1） in, This indicates the pitting risk value. Indicates the area of ​​the pit opening. The risk value contribution function, Indicates the maximum depth of the pit The risk value contribution function, Indicates the average slope of the pit The risk value contribution function, Indicates the interval distance The risk value contribution function, , , and These represent the preset weights of each contribution function.

5. The map updating method according to claim 3, characterized in that, The preset risk threshold includes: a first preset risk threshold and a second preset risk threshold that is greater than the first preset risk threshold; the step of assessing the pothole risk level of the pothole area on the road based on the relationship between the pothole risk value and the preset threshold includes: If the pothole risk value is less than the first preset risk threshold, the pothole risk level of the pothole area on the road is determined to be a low risk level. If the pothole risk value is greater than or equal to the first preset risk threshold and less than the second preset risk threshold, the pothole risk level of the pothole area on the road is determined to be a medium risk level, wherein the medium risk level is used to inform the vehicle to slow down; If the pothole risk value is greater than the second preset risk threshold, the pothole risk level of the pothole area on the road is determined to be a high risk level, wherein the high risk level is used to inform vehicles to avoid the pothole.

6. A map updating method, characterized in that, When applied to the cloud, the update method includes: The vehicle acquires road condition point clouds collected by the vehicle on the route it travels, as well as road obstacle information determined by the vehicle based on the road condition point clouds of the route it travels. The road condition point cloud is processed to obtain the road surface information of the traveled road segment, and the road surface addition and deletion information of the traveled road segment is determined based on the differential fusion result of the road surface information and the historical map in the cloud. The information on the addition and deletion of the operation boundary is obtained. The information on the addition and deletion of the operation boundary is determined by the roadside edge terminal based on the differential fusion result of the operation boundary update information and the historical map of the roadside edge terminal. The information on the update of the operation boundary is determined by the roadside edge terminal processing vehicle trajectory data and the environmental point cloud collected by the roadside edge terminal. The roadside edge terminal refers to the device installed on the roadside or on the road section. The cloud-based historical map is updated based on at least one of the road obstacle information, the road addition / deletion information, and the operation boundary addition / deletion information.

7. The map updating method according to claim 6, characterized in that, The process of processing the road condition point cloud to obtain the road surface information of the traversed road segment, and determining the addition or deletion information of the road surface of the traversed road segment based on the differential fusion result of the road surface information and the historical map in the cloud includes: The preliminary road plane in the road condition point cloud is determined based on the graph cut RANSAC algorithm. Adjust the attribution relationship between each data point in the road condition point cloud and the preliminary road plane, and determine the energy function value of each attribution relationship; The attribution relationship with the smallest energy function value is selected as the final attribution relationship; Based on the final attribution relationship, determine the in-plane data points belonging to the preliminary road plane among the various data points, and obtain the road surface information of the traversed road segment; The road surface addition and deletion information of the traveled road segment is determined based on the differential fusion results of the road surface information and the cloud historical map.

8. The map updating method according to claim 7, characterized in that, The energy function value for determining each attribution relationship includes: If the data point belongs to the preliminary road plane, the terrain point corresponding to the data point is determined in the preset mine terrain model according to the spatial location of the data point; Determine the elevation difference between the terrain point and the data point; The energy function value of the attribution relationship is determined based on the elevation difference of each data point.

9. The map updating method according to claim 8, characterized in that, The energy function value is determined based on formula (2): ；（2） in, This represents the energy function value. This represents the set of data points in the road condition point cloud. Indicates the first The attribution relationship of each data point Indicates based on the first The data items are determined by the distance between each data point and the preliminary road plane. Indicates the first The data point and the The smoothing term is generated by the different attribution relationships of the data points. Indicates the first part belonging to the preliminary road plane Each data point is based on prior terms determined by elevation difference. , These represent the weight coefficients for each item.

10. A map updating method, characterized in that, The method of updating a device installed on the roadside or on the road section through which the vehicle travels, and which is applied to the roadside edge, refers to the device installed on the roadside or on the road section through which the vehicle travels, and includes: Collect environmental point cloud data, obtain vehicle trajectory data collected by the vehicle terminal, and road obstacle information from the vehicle terminal, wherein the road obstacle information is identified by the vehicle terminal based on a visual recognition model in the collected road condition point cloud. The vehicle trajectory data and the environmental point cloud are processed to obtain the operation boundary update information of the vehicle in the frequently traveled road section, and the operation boundary addition and deletion information of the frequently traveled road section is determined based on the operation boundary update information and the differential fusion result of the historical map at the roadside edge. The road surface information of the traveled road segment is obtained by processing the road condition point cloud on the vehicle in the cloud, and the road surface addition and deletion information of the traveled road segment is determined based on the differential fusion result of the road surface information and the historical map in the cloud. The historical map of the roadside edge is updated based on at least one of the road obstacle information, the road addition / deletion information, and the operation boundary addition / deletion information.

11. The map updating method according to claim 10, characterized in that, The process of processing the vehicle trajectory data and the environmental point cloud to obtain the work boundary update information of the vehicle in frequently traveled road sections, and determining the addition or deletion information of the work boundary of the frequently traveled road sections based on the work boundary update information and the differential fusion result of the historical map at the roadside edge, includes: The vehicle trajectory data is processed using a kernel density estimation algorithm to obtain a heat map of the spatial distribution of the vehicle trajectory. Geofencing is defined based on the vehicle trajectories with heat values ​​higher than a threshold in the heat map, wherein the geofences define the frequently traveled road segments. The point cloud to be processed is determined based on the spatial location of the environmental point cloud and falls within the geofence. The road surface point cloud in the point cloud to be processed is separated based on the ray-cutting method; The operation boundary of the road surface point cloud is generated based on the Alpha Shape algorithm and used as the operation boundary update information. The addition or deletion information of the operation boundary for the frequently traveled road segment is determined based on the differential fusion results of the operation boundary update information and the historical map at the roadside edge.

12. The map updating method according to any one of claims 1-11, characterized in that, The vehicle is a mining vehicle, and the historical map is a historical map of the mine.

13. A map updating system, characterized in that, include: The vehicle-mounted terminal is used to collect vehicle trajectory data and road condition point cloud data of the vehicle-mounted terminal on the road segment it travels through; The vehicle terminal is also used to determine the road obstacle information of the traveled road segment based on the road condition point cloud, and send the road obstacle information to the roadside edge terminal and / or the cloud; The cloud is used to acquire road condition point clouds collected by the vehicle on the route, process the road condition point clouds to obtain road surface information of the route, and determine the road surface addition or deletion information of the route based on the differential fusion result of the road surface information and the historical map in the cloud. The cloud is also used to send the road surface addition / deletion information to the vehicle and / or the roadside edge. The roadside edge terminal is used to collect environmental point cloud data and vehicle trajectory data collected by the vehicle terminal, and to process the vehicle trajectory data and the environmental point cloud data to obtain the operation boundary update information of the vehicle in the frequently traveled road section. Based on the operation boundary update information and the differential fusion result of the historical map of the roadside edge terminal, the operation boundary addition and deletion information of the frequently traveled road section is determined; wherein, the roadside edge terminal refers to a device installed on the roadside or on the road section. The roadside edge end is also used to send the operation boundary addition / deletion information to the vehicle terminal and / or the cloud terminal; The vehicle end, the roadside edge end, and the cloud end all update their respective saved historical maps based on at least one of the road obstacle information, the road addition / deletion information, and the operation boundary addition / deletion information.

14. A map updating device, characterized in that, include: The processor is configured to execute the map update method according to any one of claims 1 to 12.

15. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, this instruction causes the processor to be configured to perform the map update method according to any one of claims 1 to 12.

Images