Road network generation method, device and equipment

Abstract

The application discloses a road network generation method, device and equipment. A road network generation device converts point cloud map data of a target region into open street map data, and generates Voronoi diagram data according to the open street map data. A search unit is determined according to the Voronoi diagram data, and path searching is performed according to the search unit to obtain a path searching result. The search unit represents a grid unit that meets a preset distance condition with respect to an obstacle and is a junction of multiple effective paths. The path searching result is subjected to update processing of the search unit to obtain an updated search unit. The path searching result is adjusted based on the updated search unit to obtain road network data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and particularly relates to a road network generation method, device and equipment. BACKGROUND

[0002] With the rapid development of robot technology, the demand for efficient and accurate navigation in indoor service and warehouse logistics fields is increasing. Road network (Road Network) as an abstract and compact spatial expression method can effectively support path planning and positioning of robots. Current road network extraction methods mainly include artificial design rule method, image segmentation and skeleton extraction algorithm and learning strategy based on graph optimization, but there are obvious deficiencies: the artificial rule method is low in efficiency and difficult to adapt to complex environments; the image processing method is dependent on the quality of the grid map and is prone to road network breakage or redundancy in the presence of noise and occlusion; the learning-based method requires a large amount of labeled data, has limited generalization ability, and most of the methods are for car navigation scenarios, which is difficult to meet the needs of robots in narrow channels, multi-layer structures and other special environments, resulting in poor quality and efficiency of road network generation, and the mobile robot system cannot be well adapted. Therefore, how to improve the quality and generation efficiency of the road network is a technical problem to be solved at present. SUMMARY

[0003] The embodiments of the present application provide a road network generation method, device and equipment, which can effectively improve the quality and generation efficiency of the road network.

[0004] The technical scheme of the embodiments of the present application is as follows:

[0005] In a first aspect, the embodiments of the present application provide a road network generation method, which comprises: converting point cloud map data of a target region into open street map data, and generating Voronoi diagram data according to the open street map data; determining a search unit according to the Voronoi diagram data, and performing path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit whose distance to an obstacle meets a preset distance condition and is a junction of multiple effective paths; performing update processing of the search unit on the path search result to obtain an updated search unit; and adjusting the path search result based on the updated search unit to obtain road network data.

[0006] In a second aspect, the embodiments of the present application provide a road network generation apparatus, comprising a first generation unit, a determination unit, an updating unit and a second generation unit; the first generation unit is configured to convert point cloud map data of a target region into open street map data, and generate Voronoi diagram data according to the open street map data; the determination unit is configured to determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit that meets a preset distance condition with respect to an obstacle and is a junction of multiple effective paths; the updating unit is configured to perform updating processing of the search unit on the path search result to obtain an updated search unit; and the second generation unit is configured to adjust the path search result based on the updated search unit to obtain road network data.

[0007] In a third aspect, the embodiments of the present application provide a road network generation apparatus, comprising a processor and a memory storing processor-executable instructions; when the instructions are executed by the processor, the above-mentioned road network generation method is implemented.

[0008] In a fourth aspect, the embodiments of the present application provide an electronic device, comprising a road network generation apparatus, and the electronic device is configured to perform navigation operation according to road network data generated by the road network generation apparatus.

[0009] The embodiments of the present application provide a road network generation method, apparatus and device, the road network generation apparatus can convert point cloud map data of a target region into open street map data, and generate Voronoi diagram data according to the open street map data; determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit that meets a preset distance condition with respect to an obstacle and is a junction of multiple effective paths; perform updating processing of the search unit on the path search result to obtain an updated search unit; and adjust the path search result based on the updated search unit to obtain road network data.

[0010] It can be seen that, in the generation of the road network data of the target region, the application realizes the structured extraction of environmental information through the conversion of the point cloud map to the open street map data, and provides a basic framework that conforms to the environmental topology for path generation in combination with the spatial division characteristics of the Voronoi diagram data; the searching unit ensures the safety and connectivity of the path from the root, so that even in the presence of noise interference or local occlusion in the point cloud, the path search based on the searching unit can still stably capture the effective passing nodes and avoid the road network rupture caused by environmental interference; and the searching unit as the core node of the path search can greatly reduce the spatial range of the path search and reduce unnecessary calculation consumption; the subsequent searching unit update processing further optimizes the node distribution and avoids the road network redundancy caused by redundant intersection points, so that the finally generated road network structure is more concise and can more accurately reflect the passable path characteristics of the target region. Especially for some narrow passages, multi-layer structures and other special environments, the application can conform to the passing limit of the narrow space and clearly distinguish the vertical spatial topology of the multi-layer structure through accurate node distribution and simplified path design, so as to effectively avoid the path planning failure problem that easily occurs in special environments; at the same time, the structured processing flow from data conversion to node update can avoid redundant operations, significantly improve the conversion efficiency from the original point cloud to the final road network data, and better adapt to the demand for rapid generation of road network in actual scenes; in summary, the application greatly improves the generation efficiency while ensuring the accuracy, safety and rationality of the road network. BRIEF DESCRIPTION OF DRAWINGS

[0011] Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The accompanying drawings are included only to illustrate preferred embodiments and are not to be considered as limiting of the present application.

[0012] Figure 1 Implementation flow of the road network generation method proposed by the embodiment of the application Figure One ;

[0013] Figure 2 Implementation flow of the road network generation method proposed by the embodiment of the application Figure Two ;

[0014] Figure 3 Schematic diagram of the Voronoi diagram data proposed by the embodiment of the application

[0015] Figure 4 Schematic diagram of the path search result proposed by the embodiment of the application

[0016] Figure 5 Schematic diagram of the road network data proposed by the embodiment of the application

[0017] Figure 6A schematic diagram of the component structure of the road network generation device proposed in the embodiments of the present application Figure One ;

[0018] Figure 7 A schematic diagram of the component structure of the road network generation device proposed in the embodiments of the present application Figure Two . DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. It can be understood that the specific embodiments described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that, for the convenience of description, only the parts related to the application are shown in the drawings.

[0020] With the rapid development of robot technology, indoor service robots, warehouse logistics robots and autonomous mobile devices are increasingly widely used in various fields, and the demand for efficient and accurate navigation is also increasing. Compared with traditional map representation forms such as two-dimensional grid maps and three-dimensional point cloud maps, road networks, as a more abstract and compact spatial expression carrier, can effectively support path planning and positioning tasks. In robot navigation, road networks highlight key locations and their connection relationships by modeling the environment as a set of nodes and edges, which can make path search more efficient, especially suitable for large-scale or dynamic change scenarios. In addition, the road network structure can also be better integrated with high-level task planning systems, facilitating semantic navigation and human-computer interaction. Therefore, in robot applications that emphasize real-time, flexibility and intelligence, constructing high-quality road network structures has become a key technology.

[0021] Current mainstream road network extraction methods mainly include methods based on artificial design rules, image segmentation and skeleton extraction algorithms, and learning strategies based on graph optimization. However, these methods have many problems when actually applied to robot navigation: first, artificial annotation or rule-driven methods are inefficient, not only difficult to adapt to complex and variable real environments, but also easily affected by subjective judgment; second, image processing methods usually rely on the quality of grid maps, and are unstable in the face of noise, occlusion or non-uniform resolution, which can easily lead to the generation of broken or redundant road networks; third, learning-based methods perform well in some scenarios, but often require a large amount of labeled data for training, have limited generalization ability, and have the problems of being difficult to explain and debug. More importantly, most methods focus on road network extraction in the context of car navigation, ignoring the adaptation needs of special environments such as narrow passages, multi-layer structures or indoor-outdoor mixed spaces in robot navigation, thereby limiting their applicability in mobile robot systems.

[0022] To solve the problems of low generation efficiency and poor generation quality of the current road network generation method, the embodiments of the present application are based on the data conversion and path search of Voronoi data and Open Street Map (OSM) data, which significantly improves the quality and efficiency of road network generation, and further enhances the adaptability and navigation robustness of the road network to complex scenarios. First, in terms of generation quality, the present application realizes the structured extraction of environmental information by converting point cloud map to Open Street Map data, and combines the spatial division characteristics of Voronoi data to provide a topological environment-based framework for automatic identification and connection relationship construction of key path nodes; the definition of search unit guarantees the safety and connectivity of the path from the root, and the subsequent update processing further optimizes the node distribution, avoids redundant intersections, so that the road network has good topological connectivity and geometric structure rationality, and can more accurately reflect the passable area and navigation key points in the environment, especially suitable for robot navigation needs in complex indoor and outdoor mixed scenarios; secondly, in terms of generation efficiency and automatic execution, the search unit determined based on Voronoi data and Open Street Map data greatly reduces the path search range, reduces invalid calculation and redundant operation, and significantly improves the conversion efficiency from raw data to road network; at the same time, this method greatly reduces the dependence on manual intervention through automatic node identification and connection construction, enhances the adaptability in dynamic or unknown environment, so that the road network generation can quickly respond to scenario changes and provide high-quality data support for robot path planning and behavior decision-making, effectively promoting the development of mobile robot system to a higher level of intelligence and autonomy.

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application.

[0024] An embodiment of the present application provides a road network generation method applied to a road network generation device, such as Figure 1 As shown in the figure, the road network generation method of the road network generation device can include the following steps:

[0025] Step 101, convert the point cloud map data of the target area into Open Street Map data, and generate Voronoi data according to the Open Street Map data.

[0026] In the embodiments of the present application, the road network generation device can first convert the point cloud map data of the target area into Open Street Map data, and generate Voronoi data according to the Open Street Map data.

[0027] In the embodiments of the present application, the road network generation apparatus can be an apparatus in an electronic device for generating road network data; the electronic device can further include a laser radar, for example, the electronic device can be a robot carrying a laser radar, so that the road network generation apparatus can obtain laser point cloud data collected by the laser radar carried by the robot on a target area, and then generate point cloud map data of the target area according to the laser point cloud data.

[0028] In some embodiments of the present application, data for constructing point cloud map data is not limited to being collected by a laser radar, but a camera can also be used to take or scan a target area, and image information obtained is used to construct point cloud map data; that is, the electronic device can include a camera and / or a laser radar.

[0029] In some embodiments of the present application, the road network generation apparatus can generate point cloud map data based on Simultaneous Localization and Mapping (SLAM).

[0030] Illustratively, the electronic device is a robot carrying a laser radar, and the robot can be provided with an SLAM module; the robot can move to each area in the environment of the target area, such as all rooms in a house, corridors, passageways, corners, etc. outside, to ensure that the laser radar can collect overall information of the environment; in this process, the SLAM module gradually eliminates accumulated errors through continuous positioning and map optimization, and when the robot completes a one-week traversal of the target area, i.e. returns to the starting area or covers all key areas, the SLAM module performs loop detection and optimization using the consistency of the first and last observation data, and finally outputs point cloud map data that is globally consistent and has no obvious distortion, which can be obtained by splicing a plurality of frames of laser point clouds. Assuming that the scanning frames of the laser radar from t1 to tk are represented as , the robot pose estimated by the SLAM module from t1 to tk is represented as , the point cloud map data can be represented as , wherein represents the point cloud map data finally constructed, represents the rotation matrix in the robot pose estimated by the SLAM module at the i-th moment, which is used to describe the rotation transformation relationship of the robot at the i-th moment relative to a certain initial or unified reference coordinate system, represents the translation vector in the robot pose estimated by the SLAM module at the i-th moment, represents a single laser point cloud data in the set , i.e. the coordinates of a discrete point in the environment scanned by the laser radar at the i-th moment, in addition, based on the rotation matrix and the translation vector The transformation matrix can be obtained The transformation matrix can unify rotation and translation operations through matrix multiplication to realize coordinate transformation of robot pose and point cloud map construction in SLAM.

[0031] In the embodiments of the present application, the target area can be a specific spatial range that needs to generate a road network to support robot navigation; for example, the target area is an indoor environment, such as an office building, a shopping mall, a warehouse, and the like, which is a closed or semi-closed space, or an outdoor scene, such as a park, a street, and the like, which is an open area, or a mixed indoor and outdoor space, such as a complex containing a corridor, a courtyard, and the like.

[0032] In some embodiments of the present application, when the road network generation device converts the point cloud map data of the target area into open street map data, the point cloud map data can be filtered to obtain processed map data; then the processed map data is intercepted according to preset height information to obtain intercepted point cloud map data; wherein the preset height information represents information for screening point clouds in a vertical height interval determined according to task scene requirements; finally, the first node data is determined according to the intercepted point cloud map data, and the open street map data is determined based on the first node data and the preset electronic fence data.

[0033] In the embodiments of the present application, the filtering of the point cloud map data is mainly to remove outliers and noise points in the point cloud map data to improve the quality of the point cloud map data; the specific filtering algorithm of the filtering is not limited in the present application, for example, it can be realized through targeted algorithms such as statistical filtering and radius filtering: statistical filtering analyzes the distribution density of neighbor points around each point, and determines and removes outliers that are too far away from most points; radius filtering sets a reasonable radius range and removes isolated points with insufficient neighbor points in the range; through filtering, noise points and occasional outliers caused by laser radar measurement errors such as mirror reflection and environmental interference can be accurately identified and removed, thereby reducing data redundancy, improving the smoothness and consistency of the point cloud map, and providing higher quality basic data for subsequent data conversion and road network generation.

[0034] In the embodiments of the present application, the preset height information can be a pre-set height range, and the specific value is not limited in the present application; for example, the preset height information can be a height greater than 0m and less than 1.5m from the ground.

[0035] In some embodiments of the present application, the preset height information can be set according to the actual application scenario of the electronic device; for example, the electronic device is a robot in an indoor environment, and the preset height information can be set based on the activity range of the robot. Assuming that the height of the robot is about 1.5 m, there are generally tables, chairs, garbage cans and other obstacles that affect the robot's passage in the height range from the ground to 1.5 m. Therefore, the point cloud map data in this part can be mainly focused on, that is, the preset height information is set to be greater than 0 m and less than 1.5 m. When the processed map data is intercepted according to the preset height information, the ground point cloud and the ceiling point cloud in the point cloud map data can be removed, and only the point cloud map data in the height range of 0 m to 1.5 m is retained as the intercepted point cloud map data. In addition, the preset height can also be set in combination with the actual work requirements of the robot. For example, the highest point of the reachable work range of the robot exceeds the height of the robot, which is 1.5 m, and the lowest point of the reachable work range is 0 m. Then, the upper and lower limits of the longitudinal reachable work range of the robot can be used as the basis for intercepting the point cloud map data, so as to obtain the intercepted point cloud map data.

[0036] In some embodiments of the present application, the first node data represents a set of nodes extracted from the intercepted point cloud map data and converted into an open street map format. The first node data can include spatial coordinate points that reflect key positions of the passable space of the target area, such as channel inflection points, regional connection points, and obstacle edges, after spatial feature filtering and structured processing of the intercepted point cloud map data. These nodes are stored in the format of the open street map data specification, which is an abstract expression of the core positions with navigation significance in the environment, and provides basic spatial node information for subsequent generation of Voronoi diagram and road network construction.

[0037] In an embodiment of the present application, the Voronoi diagram, also known as Dirichlet tessellation or Thiessen polygon, is a spatial division method based on a set of points.

[0038] In some embodiments of the present application, the preset electronic fence data can be polygonal region data stored in an open street map format, which is composed of discrete coordinate points with obstacle attributes to form a closed boundary.

[0039] In some embodiments of the present application, the road network generation apparatus can obtain preset electronic fence data in response to an electronic fence drawing operation. The electronic fence drawing operation can be understood as an operation related to the user setting an electronic fence for the robot received by the electronic device. The preset electronic fence data is in the form of a sequence of discrete coordinate points, each of which is assigned an "obstacle" attribute. These points are sequentially connected to form a closed polygon boundary, which intuitively outlines the profile of the area that the electronic device is prohibited from entering.

[0040] In embodiments of the present application, determining the open street map data based on the first node data and the preset electronic fence data can be fusing the first node data and the preset electronic fence data to obtain the open street map data.

[0041] In some embodiments of the present application, when fusing the first node data and the preset electronic fence data, the coordinate spaces of the first node data and the preset electronic fence data can be calibrated first, that is, the first node data and the preset electronic fence data are unified to the same coordinate system to ensure the consistency of the spatial position and lay a spatial foundation for subsequent fusion. Then, the valid nodes and their attributes in the first node data that reflect the passable area are retained as the core basis for road network generation. On the other hand, the boundary nodes and "obstacle" attributes of the preset electronic fence data are included in the overall data structure and are explicitly marked as non-crossable constraint boundaries. Conflict data can also be removed through logical verification, such as nodes in the first node data located inside the electronic fence, which are filtered because they belong to prohibited areas, ensuring that the fused data contains not only key position information in the passable environment but also the constraint range of the preset electronic fence data. Thus, the open street map data that includes "passable nodes" and "obstacle boundaries" is obtained.

[0042] In some embodiments of the present application, the road network generation apparatus can also determine selected nodes on the first node data in response to an electronic fence node selection operation, then connect the selected nodes and add "fence attributes" or "obstacle boundaries" to these connected nodes, thereby obtaining open street map data including the preset electronic fence data.

[0043] In some embodiments of the present application, when determining the first node data based on the intercepted point cloud map data, the road network generation apparatus can perform planar projection processing on the intercepted point cloud map data to obtain planar point cloud data. Then, the planar point cloud data is divided according to a preset size to obtain a plurality of grid data conforming to the preset size. Finally, point cloud aggregation processing is performed on each grid data to determine the first node data.

[0044] It should be noted that in the embodiments of the present application, after obtaining the planar point cloud data, the planar point cloud data is mainly voxelized to obtain the first node data, that is, the planar point cloud data is divided according to the preset size to obtain a plurality of grid data conforming to the preset size, and then the point cloud aggregation process of each grid data is the voxelization process.

[0045] In the embodiments of the present application, the preset size can be a side length parameter of a two-dimensional grid set according to application requirements, that is, the projection size of a voxel on a plane; the specific value of the preset size is not limited in the present application; for example, the preset size is 0.1 meters, that is, the size of each grid on the two-dimensional plane is 0.1 meters x 0.1 meters.

[0046] In the embodiments of the present application, dividing the planar point cloud data according to the preset size means that the coordinate range of the planar point cloud is taken as a boundary, and the entire planar point cloud space is uniformly divided into a plurality of square grids of equal size along the X and Y axes according to the preset size, that is, a plurality of grid data conforming to the preset size, each grid data being a two-dimensional voxel unit; then, for each grid data, all point cloud data falling into the grid can be subjected to point cloud aggregation processing, and finally only one representative point is retained for each grid as an abstract representation of the region; wherein the point cloud aggregation processing can be to retain the center point of the grid, or to calculate the average coordinates of the points in the grid, or to retain the highest or lowest point, and the specific manner is not limited in the present application.

[0047] In some embodiments of the present application, when generating the Voronoi diagram data according to the open street map data, the Delaunay triangulation can be taken as a basis, and the originally discrete point set in the open street map data can be converted into a spatial division diagram composed of Voronoi units through geometric relationship conversion.

[0048] Illustratively, first, the discrete point set in the open street map data is subjected to Delaunay triangulation, all points are connected to form a triangular grid, and each triangle needs to satisfy the "empty circle property", that is, it will not contain any other points in the point set in its circumcircle, so as to ensure that the spatial distribution of the triangular grid is uniform and conforms to the topological characteristics of the point set; then, for each point D i , all adjacent Delaunay triangles with D i as a vertex can be screened, the circumcircle centers of these triangles are calculated, and these circumcircle centers will serve as the vertices of the Voronoi diagram; then, the adjacent circumcircle centers belonging to the same D i are sequentially connected, and the polygon boundary formed thereby constitutes the D iCorresponding to the Voronoi unit, each unit represents the area range that the point "monopolizes" in space; wherein, for the point located on the convex hull of the point set, that is, the boundary of the convex polygon formed by the outermost periphery of the point set, its adjacent Delaunay triangle is only distributed on one side, so that the center of the corresponding circumscribed circle will form a Voronoi edge extending outward indefinitely after being connected, so that the Voronoi unit of such a point presents an open form.

[0049] In the embodiments of the present application, the road network generation apparatus can determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein, the search unit represents a grid cell that meets a preset distance condition with the obstacle and is a junction of multiple effective paths.

[0050] In the embodiments of the present application, the road network generation apparatus can determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein, the search unit represents a grid cell that meets a preset distance condition with the obstacle and is a junction of multiple effective paths.

[0051] In some embodiments of the present application, the search unit can be understood as a cross point far away from the obstacle.

[0052] In some embodiments of the present application, when the road network generation apparatus determines the search unit according to the Voronoi diagram data, the road network generation apparatus can determine grid map data corresponding to the Voronoi diagram data according to the Voronoi diagram data; wherein, the grid map data includes grid cells for indicating the state of each spatial position in the Voronoi diagram data; and then the first grid cell with the state of the path point in the grid map data is screened to obtain the search unit.

[0053] In the embodiments of the present application, the mesh graph data is a data form for structurally and digitally expressing the Voronoi diagram, the core of which is to accurately map the spatial features and state information of the Voronoi diagram through a fixed-resolution square mesh array; the mesh graph data is formed by a large number of small square meshes with uniform size arranged closely, the resolution of the mesh, i.e. the side length of the single small square, can be pre-set according to the environmental accuracy requirement to ensure that the local details of the Voronoi diagram can be finely described; each small mesh as a basic data unit corresponds to a specific spatial position in the Voronoi diagram and records the state of the spatial position, the state can be divided into three categories: path point, idle point and occupied point; among them, the path point corresponds to the edge of the Voronoi diagram, and its representation can be used as the boundary line of the path key component; the idle point corresponds to the area in the Voronoi diagram that is not occupied by the polygon edge or obstacle, and represents the free space that can be passed through; the occupied point corresponds to the area covered by the obstacle in the Voronoi diagram, such as walls, equipment, etc. in the original environment, and represents the obstacle space that cannot be passed through.

[0054] In the embodiments of the present application, by determining the mesh graph data corresponding to the Voronoi diagram data, the geometric features and environmental attributes of the Voronoi diagram can be converted into structured mesh data, which not only retains the advantages of the Voronoi diagram in spatial division, but also provides an intuitive and computable data basis for subsequent path search, obstacle avoidance and other operations, thereby helping to improve the generation quality and efficiency of the road network.

[0055] In some embodiments of the present application, the preset distance condition is the side length of the grid unit greater than or equal to the first value; when the road network generation device performs screening processing on the first grid unit with the state of the grid unit in the mesh graph data as a path point to obtain the search unit, the second grid unit with the distance between the first grid unit and the grid unit with the state of the occupied point less than the side length of the grid unit of the first value can be determined in the first grid unit; the second grid unit in the first grid unit is removed, and the remaining grid unit is taken as a third grid unit; the grid unit with the first neighbor grid unit with the state of the path point in the third grid unit and the number of the first neighbor grid unit greater than the second value is determined as the search unit.

[0056] It can be understood that in the embodiments of the present application, the third grid unit represents the grid unit with the distance between the grid unit with the state of the occupied point greater than or equal to the side length of the grid unit of the first value; that is, the third grid unit can be understood as the grid unit far away from the obstacle.

[0057] In the embodiments of the present application, the first neighbor grid unit represents the grid unit adjacent to any grid unit and with the state of the path point.

[0058] In the embodiments of the present application, the specific values of the first value and the second value are not limited in the present application, and the first value and the second value are both integers greater than 0; for example, the first value can be 2, and the second value can be 2.

[0059] In some embodiments of the present application, the first value can be determined based on the physical size of the electronic device and the side length of the grid unit; wherein the physical size can be the body radius of the electronic device.

[0060] In some embodiments of the present application, when the first value is determined based on the physical size of the electronic device and the side length of the grid unit, the selected integer can be determined as the first value in the case that the product of the side length of the grid unit and the selected integer is greater than the body radius of the electronic device.

[0061] For example, since the search unit is the key node of the path, it is necessary to provide sufficient passing space for the electronic device, assuming that the electronic device is a robot, the body radius of the robot is 0.4 meters, and the side length of the grid unit is 0.3 meters, when the selected integer is 1, 1 x the side length of the grid unit < the body radius of the robot, that is, the width of 1 grid unit is less than the body radius of the robot, which is easy to cause the robot to be close to the obstacle; when the selected integer is 2, the length of 2 grid units is 0.6 meters, 2 x the side length of the grid unit > the body radius of the robot, and then the first value can be determined as 2; in addition, since the length of 2 grid units can completely cover the robot body and movement error such as turning deviation, the collision risk can be more effectively avoided.

[0062] In some embodiments of the present application, the second value can be determined based on the demand of the intersection type in the application scenario of the road network generation; the second value can be the number of path branches required by the target intersection type minus 1.

[0063] Wherein, the intersection type can be determined according to the number of path branches constituting the intersection; for example, the path branches of the intersection are 3, then the intersection belongs to “T-shaped intersection”; the path branches of the intersection are 4, then the intersection belongs to “cross intersection”; the number of path branches of the intersection can be an integer greater than 3.

[0064] For example, in the actual application scenario of a certain road network, the intersection type of “cross intersection” needs to be generated, then the second value can be determined as 3, and then the grid unit with 3 neighbor grid units with path points can be determined as the search unit in the third grid unit.

[0065] Exemplarily, from all the grid cells of the path points, those grid cells with a distance to the occupied point less than 2 grid cell edge lengths are removed, and the remaining grid cells of the path points are taken as valid path points; for example, there are originally 10 grid cells of path points (first grid cells), 3 of which (second grid cells) have a distance to the occupied point less than 2 grid cell edge lengths, the 3 grid cells are removed, and the remaining 7 grid cells (third grid cells) can be taken as valid path points for further screening; the neighbor conditions of the 7 grid cells can be determined, and the number of neighbor grid cells in a state of path points (first neighbor grid cells) around each of the 7 grid cells is counted, if the number of neighbor grid cells in a state of path points of a certain grid cell exceeds 2, the certain grid cell is determined as a search unit; for example, of the 7 grid cells, 2 grid cells each have 3 neighbor grid cells in a state of path points and 4 neighbor grid cells in a state of path points, and the 2 grid cells are the search units screened finally.

[0066] In some embodiments of the present application, when the path search device performs path search according to the search unit and obtains a path search result, a starting unit can be determined in the search unit; the starting unit represents a grid cell as a starting point of path search; a path between the starting unit and a first unit is searched to obtain a first search result; the first unit is a unit in the search unit different from the starting unit; when the first unit is the last grid cell in the search unit, or when a current search duration reaches a preset duration, the path search result is determined according to the first search result; otherwise, a path between the first unit and a next unit corresponding to the first unit is searched until the next unit is the last grid cell in the search unit, or the current search duration reaches the preset duration, the path search is determined to be completed, and the path search result is obtained.

[0067] It can be understood that in the embodiments of the present application, the next grid cell searched each time is different in the process of path search, that is, the same grid cell is not searched repeatedly; in the operation process, the searched grid cell can be marked as a searched state to prevent repeated search.

[0068] In some embodiments of the present application, the starting unit can be any grid cell in the search unit.

[0069] Exemplarily, assuming that there are four search units A, B, C and D, the preset time length is 10 minutes, and the path search process is as follows: first, A is determined as the starting unit, the path between A and the first unit B is searched, and the first search result from A to B is obtained, such as the feasible path from A to B. Since B is not the last search unit (C and D follow), and the search time length has not reached 10 minutes, the search is continued. Taking B as the current unit, the path between B and the next unit C is searched, and the search result from B to C is obtained, such as the feasible path from B to C. At this time, C is still not the last unit (D follows), and the time length has not reached, so the search is continued. Taking C as the current unit, the path between C and the next unit D is searched, and the search result from C to D is obtained, such as the feasible path from C to D. Since D is the last search unit, the search is stopped. Finally, all the search results from A to B, B to C and C to D are integrated to obtain the complete path search result. Assuming that the time consumed in the path search process from C to D reaches 10 minutes, the search is stopped regardless of whether D is the last unit or not, and the search results from A to B and B to C that have been obtained are taken as the path search result.

[0070] In step 103, the path search result is subjected to search unit updating processing to obtain updated search units.

[0071] In the embodiments of the present application, after the road network generation apparatus determines the search units according to the Voronoi diagram data and performs path search according to the search units to obtain the path search result, the path search result can be subjected to search unit updating processing to obtain updated search units.

[0072] In the embodiments of the present application, the path search result can be used to indicate paths constructed based on the search units, and new intersection points can be generated between these paths. Among these intersection points, there are intersection points with close distances between each other, and there is a certain redundancy. Therefore, the path search result is subjected to search unit updating processing, which can effectively avoid the road network redundancy caused by redundant intersection points, thereby improving the generation quality of the road network.

[0073] In some embodiments of the present application, when the road network generation apparatus subjects the path search result to search unit updating processing to obtain updated search units, the to-be-updated units can be determined according to the path intersection units in the path search result. Then, the coordinate data of the to-be-updated units is averaged to obtain updated coordinate data, and the updated search units are determined according to the updated coordinate data.

[0074] In the embodiments of the present application, the path intersection unit comprises a searching unit, and a grid unit corresponding to an intersection point formed by multiple paths in the path search result, in addition to the searching unit; that is, the path intersection unit comprises the original searching unit, and a grid unit corresponding to a new intersection point generated due to intersection between paths based on the searching unit.

[0075] In the embodiments of the present application, the unit to be updated represents a grid unit in the path intersection unit with a distance less than a preset distance threshold from an adjacent unit; that is, the unit to be updated is a grid unit with a relatively close distance in the path intersection unit, and there can be multiple units to be updated; wherein the specific value of the preset distance threshold is not limited in the present application.

[0076] Exemplarily, assuming that the preset distance threshold is the edge length of 1 grid unit, and the path intersection unit has 8; wherein the distance between the path intersection unit E and its adjacent path intersection units F and G is less than the edge length of 1 grid unit, then the three path intersection units E, F and G can be used as the unit to be updated, and an average value of the coordinate data of the path intersection units E, F and G can be obtained, which is the updated coordinate data, and the grid unit determined based on the updated coordinate data is the updated searching unit, that is, the coordinate of the updated searching unit is the updated coordinate data; assuming that the grid unit determined based on the average value of the coordinate data of E, F and G is H, then H is the updated searching unit.

[0077] In some embodiments of the present application, when the average value of the coordinate data of the unit to be updated is obtained to obtain the updated coordinate data, the average value of the abscissa of the coordinate data of the unit to be updated can be obtained to obtain the updated abscissa, the average value of the ordinate of the coordinate data of the unit to be updated can be obtained to obtain the updated ordinate, and the updated coordinate data is determined according to the updated abscissa and the updated ordinate.

[0078] Exemplarily, the updated coordinate data can be calculated by the following formula:

[0079] (1)

[0080] (2)

[0081] wherein n represents the number of the unit to be updated, represents the updated abscissa, represents the updated ordinate.

[0082] Step 104, adjusting the path search result based on the updated searching unit to obtain the road network data.

[0083] In the embodiments of the present application, after the path search result is updated by the updating of the search unit, the path search result can be adjusted based on the updated search unit to obtain the road network data.

[0084] In some embodiments of the present application, when the path search result is adjusted based on the updated search unit to obtain the road network data, the path search result can be adjusted based on the paths associated with the updated search unit, so as to obtain the final road network data.

[0085] In the embodiments of the present application, the road network data can be an environment navigation structure composed of nodes and paths in the OSM format; wherein the nodes correspond to the search units in the grid map data, and the paths correspond to the paths between the search units. That is, the road network data not only retains the spatial topological features of the grid map data, but also realizes efficient expression of the passable area of the environment through structured nodes and paths.

[0086] In some embodiments of the present application, after obtaining the road network data, attribute information can be added to the road network data; for example, a node in the road network data can be added with attribute information "coffee machine".

[0087] In the embodiments of the present application, after obtaining the road network data, attribute information can be added to the road network data, which can upgrade the road network data from a pure "spatial passable framework" to a "functional and semantic combined task carrier"; and then the electronic device can directly take the node with attribute information as a task target. For example, if the electronic device is a robot, the robot can quickly locate and plan when receiving the instruction "go to the coffee machine position", and the attribute information can supplement the environmental semantics, for example, the robot can judge the priority and feasibility of the path according to attribute information such as "firefighting passage" and "narrow road section"; so as to greatly improve the practicality and decision assistance capability of the road network data.

[0088] The embodiment of the present application provides a road network generation method, point cloud map data of a target region is converted into open street map data, and Voronoi diagram data is generated according to the open street map data; a search unit is determined according to the Voronoi diagram data, and path searching is performed according to the search unit to obtain a path search result; wherein the search unit represents a grid unit that meets a preset distance condition with an obstacle and is a junction of multiple effective paths; the path search result is subjected to search unit updating processing to obtain updated search units; and the path search result is adjusted based on the updated search units to obtain road network data. As can be seen, when generating road network data of the target region, the present application realizes the structural extraction of environmental information through the conversion of point cloud maps into open street map data, and combines the spatial division characteristics of Voronoi diagram data to provide a basic framework that conforms to the environment topology for path generation; and the search unit ensures the safety and connectivity of the path from the root, that is, even in the face of noise interference or local occlusion that may exist in the point cloud, the path search based on the search unit can still stably capture effective passing nodes to avoid road network rupture caused by environmental interference; and the search unit is used as the core node of the path search, which can greatly reduce the spatial range of the path search and reduce unnecessary calculation consumption; the subsequent search unit updating processing further optimizes the node distribution to avoid road network redundancy caused by redundant junctions, so that the finally generated road network structure is more concise and can more accurately reflect the passable path characteristics of the target region, especially for some narrow passages, multi-layer structures and other special environments, the present application can conform to the passing limit of narrow space and clearly distinguish the vertical spatial topology of multi-layer structure through accurate node distribution and simplified path design, which can effectively avoid the path planning failure problem that easily occurs in special environments; at the same time, the structured processing procedure from data conversion to node updating can avoid redundant operations, significantly improve the conversion efficiency from the original point cloud to the final road network data, and better meet the demand for rapid generation of road network in actual scenes; as can be seen, the present application greatly improves the generation efficiency while ensuring the accuracy, safety and rationality of the road network.

[0089] Based on the above embodiment, in another embodiment of the present application, exemplarily, Figure 2As shown, in generating the road network data available for robot navigation, the application can first acquire the laser point cloud collected by the laser radar (step 201), then generate the point cloud map data based on the SLAM module (step 202), then convert the point cloud map data into the OSM format (step 203), then fuse the preset electronic fence data and the point cloud map data in the OSM format (step 204), then generate the Voronoi diagram data (step 205), extract the intersection points in the Voronoi diagram data (step 206), perform path search according to the intersection points (step 207), update the redundant intersection points in the path search result (step 208), and finally determine the road network data based on the updated intersection points (step 209).

[0090] In some embodiments of the application, the process of generating road network data involves two inputs and one output, where the two inputs are laser point cloud collected by the laser radar and preset electronic fence data, and the OSM module in the road network generation device can offline convert the point cloud map data constructed by the SLAM module into an OSM file, and merge it with the preset electronic fence data, then the Voronoi module in the road network generation device can generate a Voronoi diagram according to the OSM file, and extract and update the intersection points in the grid diagram of the Voronoi diagram, to generate the final navigation available road network data for output.

[0091] In some embodiments of the application, converting the point cloud map data in the Point Cloud Data (PCD) format into the OSM format can facilitate the integration of other OSM format data into the point cloud map data; the preset electronic fence data is an OSM format file, which can be a manually drawn polygonal region, and its manifestation is discrete coordinate points, and the attributes of these coordinate points are obstacles; the robot cannot cross the boundary of the electronic fence; by adding the preset electronic fence data, the effective area of the road network can be set, preventing the robot from entering unknown space, and also limiting the feasible area of the robot, and only in the feasible area, the construction of the road network is performed, improving the construction efficiency and quality of the road network.

[0092] In some embodiments of the application, the OSM module can perform filtering processing on the point cloud map data constructed by the SLAM module to filter out outliers and noise points in the point cloud map to improve the quality of the map; then remove the ground point cloud and the ceiling point cloud therein, and only keep the point cloud within a certain height range; then project the point cloud onto a two-dimensional plane, and perform voxelization processing on the plane point cloud; the first node data can be obtained after the point cloud map data is processed in the above manner.

[0093] In some embodiments of the present application, the Voronoi module can extract the boundaries of the open street map data to construct the Voronoi diagram data, and construct a grid map aligned with the coordinate system of the Voronoi diagram data; for example, the constructed Voronoi diagram data can be as shown in FIG. 11, which can include edges 111 of the Voronoi polygons, the Voronoi diagram is also called the Voronoi polygon, and the edge can also be understood as a path, i.e. a path from one region to another region; and can also include points 112 corresponding to obstacles in the region. The grid map is composed of small square grids of fixed resolution size arranged closely, and each small square grid can be used to record the state of the corresponding spatial position of the Voronoi diagram data, and the state includes path points, i.e. edges of the Voronoi polygon, free points, and occupied points. Figure 3

[0094] In some embodiments of the present application, the road network generation device can also include a junction extraction module, which can be used to check all grid cells with path points, and eliminate grid cells with a distance from the occupied point grid cell less than 2 grid cell edge lengths from all grid cells with path points, and then in the remaining grid cells, grid cells with more than 2 neighbor grid cells with path points are taken as search cells, i.e. junctions; then the path extraction module in the road network generation device can start from the first junction based on the A-start path search algorithm, search for a path until the search termination condition is met, and the search termination condition can be that all junctions are traversed or the search time reaches a preset time; in the process of searching for the path, the search cells that have participated in the search are marked as searched states to prevent repeated searches; for example, as shown in FIG. 12, the path 211 is the searched path. Figure 3

[0095] In embodiments of the present application, the algorithm for implementing path search is not limited to the A-start path search algorithm, and other gravity potential methods, ant colony optimization algorithms, and fast random tree methods can also be used to implement path search.

[0096] For example, as shown in FIG. 13, the search cells and the path searched by searching the search cells can include a plurality of search cells 312 and paths 311 between the search cells. Figure 4

[0097] In some embodiments of the present application, after obtaining the path search result, the search cells in the path search result can also be updated; the road network generation device can include an update module, which can merge adjacent search cells and calculate the average of a coordinate as the coordinate of the updated search cell after merging, and the calculation method can be as shown in the foregoing formula (1) and formula (2).

[0098] ​​​Exemplarily, the updating module can determine the path associated with the updated search unit after updating the search unit, so as to obtain the final road network data according to the updated search unit and the updated path, for example, as shown in Figure 5 the final generated road network.

[0099] In some embodiments of the present application, the final output road network can be in OSM format.

[0100] It should be noted that the general planar mobile robot, such as humanoid robot, sweeping robot, carrying robot, inspection robot, service robot, and common indoor scene working robot, such as cleaning robot, service robot, and common outdoor scene working robot, such as inspection robot, unmanned logistics vehicle, etc. are suitable for the road network generation method proposed in the embodiments of the present application.

[0101] In summary, the embodiments of the present application do not require any prior information, but rely on the sensors carried by the robot itself, such as laser radar, to complete the automatic construction of the road network, which can adapt to different indoor and outdoor scenes. For some ways of constructing road network by relying on known floor plan, when the layout of the house changes, the road network constructed by the floor plan is no longer applicable, and using grid map to generate road network, it is difficult to strictly distinguish the obstacle boundary in some special scenes, which may generate incorrect road network. Therefore, compared with the current road network generation methods which require prior information, the method of automatically generating road network using OSM data and Voronoi diagram in the present application does not require prior information, is more flexible and reliable, and is convenient for fusion of other OSM format data to cooperatively construct road network, so that the efficiency of road network construction is effectively improved.

[0102] Based on the above embodiments, in another embodiment of the present application, a road network generation device is provided, as shown in Figure 6 The road network generation device 1 can include a first generation unit 11, a determination unit 12, an updating unit 13, and a second generation unit 14.

[0103] The first generation unit 11 can be used to convert the point cloud map data of the target area into open street map data, and generate Voronoi diagram data according to the open street map data. The determination unit 12 can be used to determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit whose distance to the obstacle meets a preset distance condition and is a junction of multiple effective paths. The updating unit 13 can be used to perform search unit updating processing on the path search result to obtain an updated search unit. The second generation unit 14 can be used to adjust the path search result based on the updated search unit to obtain road network data.

[0104] In some embodiments of the present application, the determining unit 12 can be further configured to determine, according to the Voronoi diagram data, grid diagram data corresponding to the Voronoi diagram data; the grid diagram data comprises grid cells used to indicate states of respective spatial positions in the Voronoi diagram data; and perform filtering processing based on first grid cells with a state of path points in the grid diagram data, to obtain the search unit.

[0105] In some embodiments of the present application, the preset distance condition is a side length of a grid cell greater than or equal to a first numerical value; the determining unit 12 can be further configured to determine, in the first grid cell, a second grid cell with a distance to a grid cell with a state of an occupied point less than a side length of a grid cell of the first numerical value; and remove the second grid cell in the first grid cell, and take the remaining grid cells as third grid cells; and determine, as the search unit, a grid cell with a first neighbor grid cell with a state of a path point in the third grid cell, and a number of the first neighbor grid cells greater than a second numerical value.

[0106] In some embodiments of the present application, the determining unit 12 can be further configured to determine, in the search unit, a starting unit; the starting unit represents a grid cell as a starting point of path searching; and perform searching on a path between the starting unit and a first unit to obtain a first search result; the first unit is a unit different from the starting unit in the search unit; and in a case where the first unit is the last grid cell in the search unit, or a current search duration reaches a preset duration, determine a path search result according to the first search result; otherwise, continue to search on a path between the first unit and a next unit corresponding to the first unit, until the next unit is the last grid cell in the search unit, or the current search duration reaches the preset duration, to determine that the path searching is completed, and obtain the path search result.

[0107] In some embodiments of the present application, the updating unit 13 can be further configured to determine, according to a path intersection unit in the path search result, a unit to be updated; the path intersection unit comprises the search unit, and a grid cell corresponding to an intersection point formed by a plurality of paths in the path search result, except the search unit; the unit to be updated represents a grid cell with a distance to an adjacent unit less than a preset distance threshold in the path intersection unit; and take an average value of coordinate data of the unit to be updated to obtain updated coordinate data, and determine an updated search unit according to the updated coordinate data.

[0108] In some embodiments of the present application, the first generating unit 11 can also be configured to perform filtering processing on the point cloud map data to obtain processed map data, and perform intercepting processing on the processed map data according to preset height information to obtain intercepted point cloud map data, wherein the preset height information represents information for screening point clouds in a vertical height interval determined according to task scene requirements, and the first node data is determined according to the intercepted point cloud map data, and the open street map data is determined based on the first node data and the preset electronic fence data.

[0109] In some embodiments of the present application, the first generating unit 11 can also be configured to perform planar projection processing on the intercepted point cloud map data to obtain planar point cloud data, and perform division processing on the planar point cloud data according to a preset size to obtain a plurality of grid data conforming to the preset size, and perform point cloud aggregation processing on each grid data to determine the first node data.

[0110] In the embodiments of the present application, further, Figure 7 The composition structure of the road network generation device according to the embodiments of the present application is shown in Figure Two As shown in Figure 7 The road network generation device 1 according to the embodiments of the present application can also include a processor 15, a memory 16 storing executable instructions of the processor 15, and further include a communication interface 17, and a bus 18 for connecting the processor 15, the memory 16 and the communication interface 17.

[0111] In the embodiments of the present application, the processor 15 can be at least one of an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a Central Processing Unit (CPU), a controller, a microcontroller, and a microprocessor. It can be understood that the electronic device for implementing the functions of the processor described above can also be other electronic devices, and the embodiments of the present application are not limited specifically. The road network generation device 1 can further include a memory 16, which can be connected with the processor 15. The memory 16 is used to store executable program codes, which include computer operation instructions. The memory 16 can include a high-speed RAM memory and can also include a non-volatile memory, such as at least two disk memories.

[0112] In the embodiments of the present application, the bus 18 is used to connect the communication interface 17, the processor 15, and the memory 16, and to realize the mutual communication among these devices.

[0113] In the embodiments of the present application, the memory 16 is used to store instructions and data.

[0114] Further, in the embodiments of the present application, the processor 15 is used to convert the point cloud map data of the target area into open street map data, and generate Voronoi diagram data according to the open street map data; determine a search unit according to the Voronoi diagram data, and perform path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit that is an intersection point of a plurality of effective paths and whose distance to the obstacle meets a preset distance condition; perform update processing of the search unit on the path search result to obtain an updated search unit; and adjust the path search result based on the updated search unit to obtain road network data.

[0115] In practical applications, the above-mentioned memory 16 can be a volatile memory, such as a random-access memory (RAM), or a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), or a combination of the above-mentioned memories, and provides instructions and data to the processor 15.

[0116] In addition, each functional module in the embodiment can be integrated in one processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional module.

[0117] The integrated unit, if realized in the form of a software functional module and not sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiment can essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to execute all or part of the steps of the embodiment method. The aforementioned storage medium includes a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0118] The embodiment of the application provides a road network generation device for converting point cloud map data of a target region into open street map data, and generating Voronoi diagram data according to the open street map data; determining a search unit according to the Voronoi diagram data, and performing path search according to the search unit to obtain a path search result; wherein the search unit represents a grid unit that meets a preset distance condition with an obstacle and is a junction of multiple effective paths; performing search unit updating processing on the path search result to obtain an updated search unit; and adjusting the path search result based on the updated search unit to obtain road network data. As can be seen, when generating road network data of the target region, the application realizes the structural extraction of environmental information through the conversion of point cloud maps into open street map data, and combines the spatial division characteristics of Voronoi diagram data to provide a basic framework that conforms to the environment topology for path generation; and the search unit ensures the safety and connectivity of the path from the root, so that even in the face of noise interference or local occlusion that may exist in the point cloud, the path search based on the search unit can still stably capture effective passing nodes to avoid road network rupture caused by environmental interference; and taking the search unit as the core node of path search can greatly reduce the spatial range of path search and reduce unnecessary calculation consumption; the subsequent search unit updating processing further optimizes the node distribution, avoids road network redundancy caused by redundant junctions, makes the finally generated road network structure more concise, and can more accurately reflect the passable path characteristics of the target region, especially for some narrow passages, multi-layer structures and other special environments. The application can conform to the passing limit of narrow spaces and clearly distinguish the vertical spatial topology of multi-layer structures through accurate node distribution and simplified path design, which can effectively avoid the path planning failure problem that easily occurs in special environments; at the same time, the structured processing flow from data conversion to node updating can avoid redundant operations, significantly improve the conversion efficiency from the original point cloud to the final road network data, and better adapt to the demand for rapid generation of road network in actual scenarios; in summary, the application greatly improves the generation efficiency while ensuring the accuracy, safety and rationality of the road network.

[0119] Specifically, the program instructions corresponding to the road network generation method in the embodiment can be stored on a storage medium such as an optical disc, a hard disk, a U disk, etc. When the program instructions corresponding to the road network generation method in the storage medium are read by a road network generation device or executed, the following steps are included:

[0120] The point cloud map data of the target region is converted into open street map data, and Voronoi diagram data is generated according to the open street map data;

[0121] The searching unit is determined according to the Voronoi diagram data, and path searching is performed according to the searching unit to obtain a path searching result; wherein, the searching unit represents a grid unit that is a meeting point of multiple effective paths and meets a preset distance condition with the obstacle;

[0122] The path searching result is subjected to updating processing of the searching unit to obtain an updated searching unit;

[0123] The path searching result is adjusted based on the updated searching unit to obtain road network data.

[0124] In the embodiments of the present application, an electronic device can include a road network generation apparatus, which can be used to execute the aforementioned road network generation method; the electronic device can be used to perform a navigation operation according to the road network data generated by the road network generation apparatus.

[0125] Exemplarily, the electronic device can be any type of robot, and the road network generation apparatus can be deployed in the robot, so that the robot can navigate and move based on the road network data generated by the road network generation apparatus.

[0126] In some embodiments of the present application, the electronic device can further include a 3D laser radar and / or a camera; for example, the electronic device can be a humanoid robot, a sweeping robot, a patrol robot, etc. that includes a laser radar and the road network generation apparatus.

[0127] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage, etc.) containing computer-usable program code.

[0128] The present application is described with reference to the implementation flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram and the combination of the flows and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce a device that implements the functions specified in the flow or flows and / or blocks. Figure One The device that implements the functions specified in the flow or flows and / or blocks Figure One The device that implements the functions specified in the flow or flows and / or blocks

[0129] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the Figure One function specified in the flowchart or flowsheet and / or block Figure One s or blocks.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure One function specified in the flowchart or flowsheet and / or block Figure One Figure One s or blocks.

[0131] The above embodiments are only preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Any equivalent replacement or transformation made by those skilled in the art based on the present application shall fall within the protection scope of the present application.

Claims

1. A road network generation method characterized by comprising: The method comprises: Converting point cloud map data of a target area into open street map data, and generating Voronoi diagram data according to the open street map data; Determining a search unit according to the Voronoi diagram data, and performing path search according to the search unit to obtain a path search result; wherein the search unit represents a grid cell that meets a preset distance condition with an obstacle and is an intersection point of multiple effective paths; the preset distance condition is that the edge length of the grid cell is greater than or equal to a first numerical value; Performing update processing of the search unit on the path search result to obtain an updated search unit; Adjusting the path search result based on the updated search unit to obtain road network data; The method comprises: Determining grid map data corresponding to the Voronoi diagram data according to the Voronoi diagram data; wherein the grid map data comprises grid cells for indicating the state of each spatial position in the Voronoi diagram data; Determining, in a first grid cell with a state of a path point in the grid map data, a second grid cell with a distance to a grid cell with a state of an occupied point being less than the edge length of the first numerical value; Removing the second grid cell in the first grid cell, and taking the remaining grid cells as third grid cells; Determining, in the third grid cells, a grid cell with a first neighbor grid cell with a state of a path point and a number of the first neighbor grid cells being greater than a second numerical value as the search unit.

2. The road network generation method according to claim 1, characterized by, The method comprises: Determining a starting unit in the search unit; wherein the starting unit represents a grid cell as a starting point of path search; Searching a path between the starting unit and a first unit to obtain a first search result; wherein the first unit is a unit different from the starting unit in the search unit; In a case where the first unit is the last grid cell in the search unit, or a current search duration reaches a preset duration, determining the path search result according to the first search result; otherwise, continuing to search a path between the first unit and a next unit corresponding to the first unit until the next unit is the last grid cell in the search unit, or the current search duration reaches the preset duration, determining that the path search is completed to obtain the path search result.

3. The road network generation method according to claim 2, characterized by, The method comprises: Determining a to-be-updated unit according to a path intersection unit in the path search result; wherein the path intersection unit comprises the search unit, and a grid cell corresponding to an intersection point formed by multiple paths in the path search result except the search unit; the to-be-updated unit represents a grid cell in the path intersection unit with a spacing to an adjacent unit being less than a preset distance threshold; The coordinate data of the to-be-updated unit is averaged to obtain updated coordinate data, and the updated searching unit is determined according to the updated coordinate data.

4. The road network generation method according to claim 3, characterized by, The point cloud map data of the target region is converted into open street map data, including: The point cloud map data is filtered to obtain processed map data; The processed map data is intercepted according to preset height information to obtain intercepted point cloud map data; wherein the preset height information represents information for filtering point clouds in a vertical height interval determined according to task scene requirements; First node data is determined according to the intercepted point cloud map data, and the open street map data is determined based on the first node data and preset electronic fence data.

5. The road network generation method according to claim 4, characterized by, The first node data is determined according to the intercepted point cloud map data, including: The intercepted point cloud map data is subjected to planar projection processing to obtain planar point cloud data; The planar point cloud data is divided according to a preset size to obtain a plurality of grid data conforming to the preset size; Point cloud aggregation processing is performed on each grid data to determine the first node data.

6. A road network generation apparatus characterized by comprising: The road network generation device includes a first generation unit, a determination unit, an update unit, and a second generation unit; The first generation unit is configured to convert point cloud map data of a target region into open street map data, and generate Voronoi diagram data according to the open street map data; The determination unit is configured to determine a searching unit according to the Voronoi diagram data, and perform path search according to the searching unit to obtain a path search result; wherein the searching unit represents a grid unit that meets a preset distance condition with an obstacle and is an intersection of a plurality of effective paths; the preset distance condition is that the edge length of the grid unit is greater than or equal to a first value; The update unit is configured to perform searching unit update processing on the path search result to obtain an updated searching unit; The second generation unit is configured to adjust the path search result based on the updated searching unit to obtain road network data; The determination unit is further configured to determine grid map data corresponding to the Voronoi diagram data according to the Voronoi diagram data; wherein the grid map data includes grid units for indicating the state of each spatial position in the Voronoi diagram data; in a first grid unit with a path point state in the grid map data, a second grid unit with a distance less than the edge length of the grid unit between the grid unit with an occupied point state is determined; the second grid unit in the first grid unit is removed, and the remaining grid unit is taken as a third grid unit; a grid unit with a first neighbor grid unit having a path point state in the third grid unit and a number of the first neighbor grid units greater than a second value is determined as the searching unit.

7. A road network generation apparatus characterized by comprising: The road network generation device includes a processor and a memory storing processor-executable instructions; when the instructions are executed by the processor, the method of any one of claims 1 to 5 is implemented.

8. An electronic device, comprising: The electronic device comprises the road network generation apparatus as claimed in claim 7. The electronic device is configured to perform a navigation operation according to the road network data generated by the road network generation apparatus.

Images