Road network construction method and apparatus, and electronic device

Abstract

Provided in the present application are a road network construction method and apparatus, and an electronic device. The method comprises: acquiring crowdsourced trajectory information and environment sensing information; processing the crowdsourced trajectory information in layers on the basis of a plurality of preset height ranges, so as to obtain a crowdsourced trajectory corresponding to each preset height range; for each preset height range, constructing a corresponding region image on the basis of the corresponding crowdsourced trajectory and corresponding environment sensing information; performing image recognition on the region image, so as to obtain a regional road network within the preset height range; and connecting regional road networks within the plurality of preset height ranges, so as to obtain a target road network. In the method provided in the present application, by means of processing crowdsourced trajectory data in layers, a regional road network in each layer can be determined on the basis of crowdsourced trajectory data and environment sensing information that correspond to each layer, and then a road network is obtained by means of connecting regional road networks in layers, thereby improving the accuracy of the road network.

Description

Road network construction method, device and electronic equipment

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on December 29, 2023, with application number 2023118681497 and application name “Road Network Construction Method, Device and Electronic Equipment”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of data processing technology, and specifically to a road network construction method, device and electronic equipment. Background Art

[0003] Road networks, also known as road networks, are the foundation for various transportation functions, such as vehicle navigation and route planning. Building these networks requires a large amount of road data as input. In related technologies, the data used to construct these networks primarily comes from two sources: satellite imagery and crowdsourced trajectory data uploaded by vehicles.

[0004] However, since satellite images can only provide planar images, they cannot be collected in partially obscured scenes (such as tunnels, elevated roads, etc.), and crowdsourced trajectory data is prone to data accumulation in obscured areas, making road network construction difficult. Moreover, crowdsourced trajectory data is easily affected by the driver's subjective driving behavior, which can easily cause the obtained road network to deviate from the actual road, resulting in the road network obtained based on satellite images or crowdsourced trajectories being inaccurate.

[0005] Summary of the Invention

[0006] In order to solve or partially solve the problems existing in the related art, the embodiments of the present application propose a road network construction method, device and electronic equipment, which can perform highly layered processing of crowdsourced trajectories, avoiding the accumulation of crowdsourced trajectory data in the height direction; at the same time, combining environmental perception information and crowdsourced trajectory data to obtain a road network, avoiding the problem of crowdsourced trajectories deviating from the road, thereby improving the accuracy of the constructed road network.

[0007] The embodiments of the present application are implemented using the following technical solutions:

[0008] In a first aspect, an embodiment of the present application provides a road network construction method, which includes: obtaining crowdsourced trajectory information and environmental perception information uploaded by a target device, the crowdsourced trajectory information including trajectory point information at multiple times, each trajectory point information including the position and height of the trajectory point, and the environmental perception information including the environmental information collected by the target device at each trajectory point; performing layered processing on the crowdsourced trajectory information based on multiple preset height ranges to obtain crowdsourced trajectories corresponding to each preset height range; for each preset height range, based on the crowdsourced trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory, constructing an area image corresponding to the preset height range; performing image recognition on the area image corresponding to the preset height range to obtain a regional road network for the preset height range; and connecting the regional road networks of multiple preset height ranges to obtain a target road network.

[0009] In a second aspect, an embodiment of the present application provides a road network construction device, which includes: an acquisition module for acquiring crowdsourced trajectory information and environmental perception information uploaded by a target device, the crowdsourced trajectory information including trajectory point information at multiple moments, each trajectory point information including the position and height of the trajectory point, and the environmental perception information including the environmental information collected by the target device at each trajectory point; a layering module for performing layered processing on the crowdsourced trajectory information based on multiple preset height ranges to obtain crowdsourced trajectories corresponding to each preset height range; a conversion module for constructing, for each preset height range, an area image corresponding to the preset height range based on the crowdsourced trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory; a processing module for performing image recognition on the area image corresponding to the preset height range to obtain the area road network of the preset height range; and an output module for connecting the area road networks of multiple preset height ranges to obtain the target road network.

[0010] In some embodiments, the processing module further includes a segmentation unit, a post-processing unit, and a combination unit. The segmentation unit is configured to perform instance segmentation on the regional image within a preset height range to obtain road trunk image blocks and intersection image blocks; the post-processing unit is configured to extract centerlines based on the road trunk image blocks to obtain road trunk centerlines; and to perform polygon fitting based on the intersection image blocks to obtain intersection polygons; and the combination unit is configured to obtain the regional road network within the preset height range based on the road trunk centerlines and intersection polygons.

[0011] In some embodiments, the regional road network includes the traffic attribute information of the road trunk, and the combination unit is further used to determine the traffic attribute information of the road trunk based on the trajectory direction and direction indication information of the crowdsourced trajectory in the regional image, and the trajectory direction of the crowdsourced trajectory is determined based on the collection time corresponding to each trajectory point in the crowdsourced trajectory; based on the center line of the road trunk, the intersection polygon, and the traffic attribute information of the road trunk, the regional road network of the preset height range is obtained.

[0012] In some embodiments, the regional road network includes multiple trunks, and the road network construction device also includes a storage module, which is used to determine the predecessor and successor relationship of each trunk based on the intersection polygon of the regional road network and the traffic attribute information of each trunk connected to the intersection polygon, and the predecessor and successor relationship is used to characterize the connection status of the trunk and other trunks; the predecessor and successor relationship of the trunk is associated with the trunk and stored.

[0013] In some embodiments, the layering module is further used to determine the preset height range to which each trajectory point belongs based on the height of each trajectory point in the crowdsourced trajectory information; and to connect multiple trajectory points within each preset height range based on the collection time corresponding to each trajectory point to obtain the crowdsourced trajectory corresponding to each preset height range.

[0014] In some embodiments, multiple preset height ranges are obtained based on multiple preset height values, and the output module is also used to determine the logical connectivity relationship between the regional road networks based on the preset height range corresponding to each regional road network and the preset height value corresponding to each preset height range; and connect the regional road networks of multiple preset height ranges based on the logical connectivity relationship to obtain the target road network.

[0015] In some embodiments, the processing module is also used to project each track point in the crowdsourcing track corresponding to the preset height range to a preset grid; based on the position of each track point in the crowdsourcing track corresponding to the preset height range, the environmental perception information corresponding to each track point is projected to the preset grid to obtain an area image corresponding to the preset height range.

[0016] In a third aspect, an embodiment of the present application provides an electronic device comprising: one or more processors; a memory; and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the above method.

[0017] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, which stores program code, and the program code can be called by a processor to execute the above method.

[0018] A road network construction method provided in an embodiment of the present application includes: obtaining crowdsourced trajectory information and environmental perception information uploaded by a target device; performing layered processing on the crowdsourced trajectory information based on multiple preset height ranges to obtain crowdsourced trajectories corresponding to each preset height range; for each preset height range, constructing a regional image corresponding to the preset height range based on the crowdsourced trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory; performing image recognition on the regional image corresponding to the preset height range to obtain a regional road network for the preset height range; and connecting the regional road networks of multiple preset height ranges to obtain a target road network. The road network construction method provided in the present application utilizes multiple preset height ranges to perform layered processing on crowdsourced trajectory data, so that the regional road network of each layer can be determined based on the crowdsourced trajectory data and environmental perception information corresponding to each layer, and then the regional road networks of each layer are connected to obtain a road network, thereby avoiding the problem of data accumulation of crowdsourced trajectory data at different heights, which makes road network construction difficult; at the same time, since the regional road network is obtained based on the crowdsourced trajectory and environmental perception information corresponding to each layer, even if the crowdsourced trajectory deviates from the actual road, an accurate road network can be obtained by combining the environmental perception information and the crowdsourced trajectory data, thereby avoiding the problem of the road network deviating from the actual road due to the deviation of the crowdsourced trajectory from the actual road, thereby improving the accuracy of the road network.

[0019] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The above and other objects, features and advantages of the present application will become more apparent through a more detailed description of exemplary embodiments of the present application in conjunction with the accompanying drawings, wherein the same reference numerals generally represent the same components in the exemplary embodiments of the present application.

[0021] FIG1 is a schematic diagram showing a flow chart of a road network construction method provided in an embodiment of the present application;

[0022] FIG2 shows a schematic flow chart of step S120 in FIG1 according to an embodiment of the present application;

[0023] FIG3 shows a schematic flow chart of step S130 in FIG1 according to an embodiment of the present application;

[0024] FIG4 shows a schematic flow chart of step S140 in FIG1 according to an embodiment of the present application;

[0025] FIG5 shows a schematic flow chart of step S144 in FIG4 provided in an embodiment of the present application;

[0026] FIG6 shows another flow chart of the road network construction method provided in an embodiment of the present application;

[0027] FIG7 shows a schematic flow chart of step S150 in FIG1 according to an embodiment of the present application;

[0028] FIG8 shows a schematic diagram of a road network construction device provided in an embodiment of the present application;

[0029] FIG9 shows a schematic structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] The following describes embodiments of the present application in more detail with reference to the accompanying drawings. Although the accompanying drawings illustrate embodiments of the present application, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. Rather, these embodiments are provided to make the present application more thorough and complete, and to fully convey the scope of the present application to those skilled in the art.

[0031] The terms used in this application are for the purpose of describing specific embodiments only and are not intended to limit this application. As used in this application and the appended claims, the singular forms "a," "an," "the," and "the" are intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0032] It should be understood that although the terms "first", "second", "third", etc. may be used in this application to describe various information, this information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of this application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0033] A road network is a network structure composed of nodes and edges, which is used to represent a transportation network. It is an effective transportation network model that can clearly present the basic characteristics of a transportation network and is widely used in transportation planning, logistics and transportation, and other fields.

[0034] In related technologies, road network construction methods can be divided into two main categories according to the different data sources. One is to extract road networks based on satellite images, such as image processing of satellite images, or inference analysis of satellite images based on deep learning models. However, satellite images can only provide 2D plane information and cannot provide effective information for scenes such as tunnels, elevated roads, cross-floor roads, and physical occlusions. The other is to extract road networks based on crowdsourced trajectories. However, crowdsourced trajectories are greatly affected by the driver's subjective driving behavior and are prone to deviate from the actual road, making the extracted road network inaccurate.

[0035] In order to solve the above problems, the present application provides a road network construction method, which includes: obtaining crowdsourced trajectory information and environmental perception information uploaded by a target device, the crowdsourced trajectory information includes trajectory point information at multiple times, each trajectory point information includes the position and height of the trajectory point, and the environmental perception information includes the environmental information collected by the target device at each trajectory point; based on multiple preset height ranges, the crowdsourced trajectory information is layered to obtain crowdsourced trajectories corresponding to each preset height range; for each preset height range, based on the crowdsourced trajectory corresponding to the preset height range, and the environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory, an area image corresponding to the preset height range is constructed; image recognition is performed on the area image corresponding to the preset height range to obtain a regional road network of the preset height range; and the regional road networks of multiple preset height ranges are connected to obtain a target road network.

[0036] The road network construction method provided by the present application utilizes multiple preset height ranges to perform layered processing on crowdsourced trajectory data, so that the regional road network of each layer can be determined based on the crowdsourced trajectory data and environmental perception information corresponding to each layer, and then the regional road networks of each layer are connected to obtain a road network, thereby avoiding the problem of data accumulation of crowdsourced trajectory data at different heights, which leads to difficulties in road network construction; at the same time, since the regional road network is obtained based on the crowdsourced trajectory and environmental perception information corresponding to each layer, even if the crowdsourced trajectory deviates from the actual road, an accurate road network can be obtained by combining the environmental perception information and the crowdsourced trajectory data, thereby avoiding the problem of the road network deviating from the actual road due to the deviation of the crowdsourced trajectory from the actual road, thereby improving the accuracy of the road network.

[0037] The embodiments provided in this application will be described below with reference to the accompanying drawings.

[0038] Please refer to FIG1 , which shows a flow chart of a road network construction method provided in an embodiment of the present application. The road network construction method includes steps S110 to S150:

[0039] S110: Obtain crowdsourced trajectory information and environmental perception information uploaded by the target device.

[0040] The target device may be a vehicle, a mobile robot, a drone, or other mobile device that can provide the required information; further, the target device may be one or more.

[0041] For each target device, its crowdsourced trajectory information includes trajectory point information at multiple moments. Each trajectory point information includes the position and height of the trajectory point. The environmental perception information includes the environmental information collected by the target device at each trajectory point, such as lane lines, road edges, parking spaces, and road signs in the environment.

[0042] In some embodiments, environmental perception information may include image information and point cloud information. For example, the image information may be used to identify a target object, such as a lane line or road sign, and the corresponding point cloud information may be used to determine the location of the target object, such as the distance between the lane line and the target device, or the distance between the road sign and the target device.

[0043] It is worth mentioning that the target device is installed with various collection devices for collecting trajectory information and environmental perception information, such as cameras, radars, position sensors, height sensors, etc.; further, since the trajectory information and environmental perception information are collected by different collection devices, in order to determine the correspondence between the trajectory information and the environmental perception information, after the target device collects the required information, it can associate the information according to the collection timestamps of each information, and then package and upload it.

[0044] S120 , performing layered processing on the crowdsourcing trajectory information based on multiple preset height ranges to obtain crowdsourcing trajectories corresponding to each preset height range.

[0045] Among them, the preset height range can be determined based on the actual application scenario; for example, for a multi-story underground parking lot, the floor height of each floor is m, then the preset height ranges can be (0, -m), (-m, -2m), (-2m, -3m), etc.; of course, if the preset height ranges can also be inconsistent, for example, (0, m), (m, n), (n, p), etc., where m<n<p, and both are positive; the preset height range can be set according to actual needs, and no specific limitation is made here.

[0046] It should be noted that the crowdsourcing trajectory corresponding to each preset height range may include crowdsourcing trajectories corresponding to multiple scenes within the preset height range. For example, for the preset height range (m, n), it corresponds to the range of the negative first floor of the parking lot. The crowdsourcing trajectory information belonging to the preset height range includes multiple trajectory points, among which N1 trajectory points belong to the negative first floor of parking lot A, and N2 trajectory points belong to the negative first floor of parking lot B. It is necessary to separately determine the crowdsourcing trajectory corresponding to parking lot A within the preset height range, and the crowdsourcing trajectory corresponding to parking lot B within the preset height range. Furthermore, the specific scene to which the trajectory point belongs can be determined based on the position of the trajectory point. For example, based on the position of the trajectory point, the distance between each trajectory point within the preset height range is calculated. If the distance between the trajectory points is less than the preset distance, it is considered that the corresponding two trajectory points belong to the same scene.

[0047] It is understandable that the crowdsourced trajectory information is just some isolated trajectory point information. To obtain the crowdsourced trajectory corresponding to the preset height range, it is necessary to connect the trajectory points according to the collection time of each trajectory point. In some embodiments, please refer to Figure 2, which shows a flow diagram of step S120 in Figure 1 provided in an embodiment of the present application. Step S120 includes steps S121-S122:

[0048] S121 : Determine a preset height range to which each trajectory point belongs based on the height of each trajectory point in the crowdsourced trajectory information.

[0049] It is understandable that for the crowdsourced trajectory information uploaded by the same target device, after mapping each trajectory point to its corresponding preset height range according to its height, the problem of data accumulation at different heights of the trajectory points can be avoided.

[0050] S122 . Connect multiple trajectory points within each preset height range based on the collection time corresponding to each trajectory point to obtain a crowdsourced trajectory corresponding to each preset height range.

[0051] For example, according to the collection time corresponding to each trajectory point, each trajectory point is connected in sequence according to the order of the collection time, so as to obtain the crowdsourcing trajectory corresponding to each preset height range.

[0052] It is worth mentioning that since the received crowdsourced trajectory information may be crowdsourced trajectory information uploaded by multiple target devices respectively, when determining the crowdsourced trajectory, it is necessary to confirm the trajectory corresponding to each target device separately, so as to obtain the crowdsourced trajectory corresponding to the preset height range based on the trajectories of multiple target devices.

[0053] Through the method provided in the embodiment of the present application, the crowdsourcing trajectory information is divided into different preset height ranges, so that trajectory points with similar positions but different heights on the 2D plane can be distinguished, avoiding the stacking of data at different heights on the 2D plane, and facilitating the acquisition of accurate crowdsourcing trajectories.

[0054] S130 . For each preset height range, construct an area image corresponding to the preset height range based on a crowdsourced trajectory corresponding to the preset height range and environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory.

[0055] It should be noted that after the trajectory points are connected according to the collection time of the trajectory points to obtain the crowdsourced trajectory, the crowdsourced trajectory can be converted to a plane according to the position of each trajectory point; further, since the environmental perception information is one-to-one corresponding to the trajectory point, the environmental perception information corresponding to each trajectory point in the crowdsourced trajectory can also be converted to the same plane, thereby obtaining the regional image corresponding to the preset height range.

[0056] In some embodiments, the environmental perception information is point cloud information, and a regional image is obtained based on the crowdsourcing trajectory and the environmental perception information, that is, the point cloud data is converted into image data; in related technologies, rasterization processing can be used to divide the point cloud data into several grids according to certain rules, and then the point cloud data in each grid is converted into pixel values ​​to form an image; or a projection method can be used to project the point cloud data onto a target two-dimensional plane to form an image; or the point cloud data can be rendered based on a deep learning network to generate an image, etc. The specific method is not limited here.

[0057] For example, in some embodiments, taking the environmental perception information as point cloud information, rasterization is performed on the environmental perception information and crowdsourced trajectories to obtain a regional image. Referring to FIG3 , FIG3 shows a flow diagram of step S130 in FIG1 provided in an embodiment of the present application. Step S130 includes steps S131-S132:

[0058] S131 , projecting each track point in the crowdsourced track corresponding to the preset height range to a preset grid.

[0059] The projection can be understood as coordinate conversion, that is, the trajectory point is converted to the coordinate system of the preset grid according to the coordinate conversion relationship between the coordinate system of the trajectory point and the coordinate system of the preset grid.

[0060] S132 . Based on the position of each track point in the crowdsourced track corresponding to the preset height range, project the environmental perception information corresponding to each track point onto a preset grid to obtain an area image corresponding to the preset height range.

[0061] It can be understood that since the environmental perception information is collected by the target device at the trajectory point; the relative position relationship between the environmental perception information and the corresponding trajectory point can be determined through the environmental perception information; therefore, after the trajectory point is projected to the preset grid, the corresponding environmental perception information can be projected into the preset grid based on the relative position relationship between the environmental perception information and the corresponding trajectory point.

[0062] Furthermore, after projecting the environmental perception information corresponding to each trajectory point into a preset grid, the point cloud data in each grid is converted into pixel values, thereby obtaining an area image corresponding to the preset height range.

[0063] It is understandable that in the above-mentioned process of processing environmental perception information, the types of information will not be reduced; for example, the obtained regional image contains both crowdsourced trajectories and various environmental perception information, such as lane lines, parking spaces, road signs, etc.

[0064] It is worth mentioning that the size of each grid in the preset grid determines the resolution of the regional image. The smaller the grid, the higher the resolution, but the processing efficiency and resources required are greater. Furthermore, the size of the preset grid can be manually determined, such as a preset fixed value; it can also be determined based on the scene corresponding to the crowdsourcing trajectory. For example, after layering the crowdsourcing trajectory information, for crowdsourcing trajectories corresponding to negative numbers in the preset height range, it may be an underground scene with a small scene range and more environmental perception information such as lane lines and parking spaces. In this case, a smaller grid can be used to obtain a higher resolution; for crowdsourcing trajectories corresponding to positive numbers greater than 0 in the preset height range, it may be an elevated scene with a large scene range and a simple scene structure. In this case, a larger grid can be used to obtain faster processing efficiency.

[0065] S140 , performing image recognition on an area image corresponding to a preset height range to obtain an area road network within the preset height range.

[0066] Among them, image recognition refers to the processing, analysis and understanding of images to identify various different patterns of targets and objects in the images.

[0067] In related image recognition technologies, image recognition can be implemented in a variety of ways, such as semantic segmentation, target detection, instance segmentation, etc.; further, the various implementation methods of image recognition can be specifically executed through pre-trained deep learning models, or implemented by corresponding image recognition algorithms. The specific method of image recognition can be selected according to actual needs and is not specifically restricted here.

[0068] In some embodiments, image recognition is performed on the regional image in an instance segmentation manner. Please refer to FIG4 , which shows a flow chart of step S140 in FIG1 provided in an embodiment of the present application. Step S140 includes steps S141-S144:

[0069] S141 , performing instance segmentation on the regional image within a preset height range to obtain road trunk image blocks and road junction image blocks.

[0070] Among them, instance segmentation is to mark each object instance in the image and accurately mark the boundary of each object.

[0071] In some embodiments, instance segmentation can be implemented by a pre-trained deep learning model; wherein the pre-trained deep learning model can be trained by multiple manually labeled sample images.

[0072] S142 , extracting a center line based on the road trunk image block to obtain a road trunk center line.

[0073] Among them, centerline extraction refers to extracting the main axis of an object.

[0074] In related technologies, centerline extraction can be performed using image processing software based on morphological methods (such as skeleton extraction in open CV), or can be implemented using image processing algorithms such as Hough Transform, without specific limitation here.

[0075] S143 . Perform polygon fitting based on the intersection image block to obtain an intersection polygon.

[0076] Among them, polygon fitting is a common image processing technology that can fit multiple points into a polygon and is often used to detect edges and contours in images.

[0077] In related technologies, polygon fitting can be implemented by various polygon fitting algorithms, such as the least squares method, the Ramer-Douglas-Peucker algorithm (RDP algorithm), the convex hull algorithm, and the like.

[0078] S144. Obtain a regional road network within the preset height range based on the road trunk centerline and the intersection polygon.

[0079] For example, the road trunk centerline and the intersection polygon are connected to obtain the regional road network within the preset height range; wherein the connection order can be determined according to the positional relationship between the road trunk image block and the intersection image block in the regional image.

[0080] It is worth mentioning that since the road center lines and intersection polygons are extracted based on regional images, in the process of connecting the road center lines and intersection polygons to obtain the regional road network, the road center lines and intersection polygons need to be converted into the coordinate system required by the road network before connection.

[0081] It is understandable that by extracting the center line of the road trunk image block and performing polygon fitting on the intersection image block, problems such as overlap and breakage between the road trunk image block and the intersection image block can be avoided, thereby facilitating the construction of the road network.

[0082] In some embodiments, the regional road network also includes traffic attribute information of the road trunk. Please refer to Figure 5, which shows a flow diagram of step S144 in Figure 4 provided in an embodiment of the present application. Step S144 includes steps S1441-S1442:

[0083] S1441: Determine the traffic attribute information of the road based on the trajectory direction and direction indication information of the crowdsourced trajectory in the regional image.

[0084] The trajectory direction of the crowdsourcing trajectory is determined based on the collection time corresponding to each trajectory point in the crowdsourcing trajectory; for example, the trajectory direction of the crowdsourcing trajectory is from the trajectory point with an earlier collection time to the trajectory point with a later collection time.

[0085] The direction indication information is determined from the environmental perception information and can indicate the direction of road travel, such as road signs, guide lane lines, etc.

[0086] It can be understood that the traffic attribute information of a road trunk mainly includes the trafficable directions of the road trunk, such as straight, left turn, right turn, one-way traffic, two-way traffic, etc.; for example, if there are two trajectories in opposite directions on the same road trunk at the same time, it can be determined that the road trunk is two-way traffic; if the direction indication information is identified as a one-way traffic sign, it can be determined that the road trunk is one-way traffic.

[0087] S1442. Based on the centerline of the road trunk, the intersection polygon, and the traffic attribute information of the road trunk, obtain the regional road network of the preset height range.

[0088] Specifically, after connecting the center lines of the road trunks and the intersection polygons, the traffic attribute information of the road trunks is associated with the corresponding road trunks to obtain the regional road network of the preset height range.

[0089] In some implementations, given that there may be multiple road trunks in a regional road network, in practical application scenarios of the road network, such as navigation scenarios, not only is the traffic attribute information of each road trunk required, but also the logical connectivity relationship between multiple road trunks needs to be considered simultaneously to achieve better navigation planning. Therefore, please refer to Figure 6, which shows another flowchart of the road network construction method provided in an embodiment of the present application. After step S1442, the road network construction method may also include steps S1442a-S1442b:

[0090] S1442a. Based on the intersection polygons of the regional road network and the traffic attribute information of the trunk roads connected to the intersection polygons, determine the predecessor and successor relationship of the trunk roads.

[0091] Among them, the predecessor and successor relationship is used to characterize the connection between the trunk and other trunks, specifically, the connection relationship between the trunk and the previous trunk and the next trunk respectively; for example, for trunk A, its previous trunk is B and the next trunk is C. If trunk B can travel to trunk A and trunk A can travel to trunk C, then the predecessor of trunk A is trunk B and the successor is trunk C, which means that path BAC can be planned during navigation; for another example, for trunk A, the previous trunks physically connected to it are B1 and B2, and the next trunk physically connected to it is C, but the traffic attribute information of trunk B1 is left turn, and it cannot enter trunk A, that is, although trunk B1 and trunk A are physically connected, it is impossible to enter trunk A from trunk B1 during actual driving, then the predecessor of trunk A is trunk B2 and the successor is trunk C, which means that path B2-AC can be planned during navigation.

[0092] S1442b. Associate the predecessor and successor relationship of the road trunk with the road trunk and store it.

[0093] By associating and storing road trunks with their predecessor-successor relationships, in actual application scenarios, the corresponding predecessor-successor relationships can be called according to the road trunks, thereby improving the practicality of the road network; for example, in navigation scenarios, advance path planning can be achieved based on the predecessor-successor relationships of the road trunks.

[0094] S150 , connecting regional road networks within multiple preset height ranges to obtain a target road network.

[0095] It can be understood that the logical connection relationship between multiple preset height ranges can be directly determined based on the specific height values; further, the logical connection relationship between multiple regional road networks can be determined based on the logical connection relationship between different preset height ranges.

[0096] For example, in some embodiments, multiple preset height ranges are obtained based on multiple preset height values. For example, if the preset height values ​​are m, n, and p, the corresponding preset height ranges may be (0, m), (m, n), and (n, p). Please refer to FIG7 , which shows a flow diagram of step S150 in FIG1 provided in an embodiment of the present application. Step S150 includes steps S151-S152:

[0097] S151. Determine the logical connectivity relationship between the regional road networks based on the preset altitude range corresponding to each regional road network and the preset altitude values ​​corresponding to each preset altitude range.

[0098] Obviously, if two preset height ranges include the same preset height value, then the two preset height ranges have a logical connection relationship; if the two preset height ranges do not include the same preset height value, then the two preset height ranges do not have a logical connection relationship.

[0099] S152. Connect regional road networks within multiple preset height ranges based on a logical connectivity relationship to obtain a target road network.

[0100] For example, the end point of a regional road network within a preset altitude range may be connected to the start point of a regional road network within another preset altitude range having a logical connectivity relationship.

[0101] It can be understood that since the regional road network of each preset height range may include regional road networks under multiple scenarios, such as for the preset height range (0, m), its regional road network may include the regional road network of parking lot A and the regional road network of parking lot B, the regional road networks of multiple preset height ranges are connected based on the logical connectivity relationship. The regional road networks of multiple preset height ranges in each scenario are connected separately. By connecting the regional road networks of multiple preset height ranges in the same scenario, the road network in the scenario can be obtained.

[0102] The road network construction method provided in the present application obtains crowdsourced trajectory information and environmental perception information uploaded by a target device; performs layered processing on the crowdsourced trajectory information based on multiple preset height ranges to obtain crowdsourced trajectories corresponding to each preset height range; for each preset height range, constructs a regional image corresponding to the preset height range based on the crowdsourced trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory; performs image recognition on the regional image corresponding to the preset height range to obtain a regional road network for the preset height range; and connects the regional road networks of multiple preset height ranges to obtain a target road network. Through the road network construction method provided by the present application, the crowdsourced trajectory data is layered and processed using multiple preset height ranges, so that the regional road network of each layer can be determined based on the crowdsourced trajectory data and environmental perception information corresponding to each layer, and then the regional road networks of each layer are connected to obtain a road network, avoiding the problem of data accumulation of crowdsourced trajectory data at different heights, which leads to difficulties in road network construction; at the same time, since the regional road network is obtained based on the crowdsourced trajectory and environmental perception information corresponding to each layer, even if the crowdsourced trajectory deviates from the actual road, an accurate road network can be obtained by combining the environmental perception information and the crowdsourced trajectory data, avoiding the problem of the road network deviating from the actual road due to the deviation of the crowdsourced trajectory from the actual road, thereby improving the accuracy of the road network.

[0103] Based on the road network construction method provided in the aforementioned embodiment, the present application also provides a road network construction device.

[0104] Please refer to FIG8 , which shows a schematic diagram of a road network construction device provided in an embodiment of the present application. The road network construction device 200 includes:

[0105] Acquisition module 210, for acquiring crowdsourced trajectory information and environmental perception information uploaded by the target device. The crowdsourced trajectory information includes trajectory point information at multiple moments, each trajectory point information including the location and altitude of the trajectory point. The environmental perception information includes the environmental information collected by the target device at each trajectory point.

[0106] A stratification module 220 is configured to perform stratification processing on the crowdsourced trajectory information based on a plurality of preset height ranges to obtain crowdsourced trajectories corresponding to each preset height range;

[0107] A conversion module 230 is configured to construct, for each preset height range, an image of the area corresponding to the preset height range based on a crowdsourced trajectory corresponding to the preset height range and environmental perception information corresponding to multiple trajectory points in the crowdsourced trajectory;

[0108] The processing module 240 is used to perform image recognition on the regional image corresponding to the preset height range to obtain the regional road network of the preset height range;

[0109] The output module 250 is used to connect the regional road networks of multiple preset height ranges to obtain a target road network.

[0110] In some embodiments, the processing module 240 further includes a segmentation unit, a post-processing unit, and a combination unit. The segmentation unit is configured to perform instance segmentation on the regional image within a preset height range to obtain road trunk image blocks and intersection image blocks; the post-processing unit is configured to extract centerlines based on the road trunk image blocks to obtain road trunk centerlines; and to perform polygon fitting based on the intersection image blocks to obtain intersection polygons; and the combination unit is configured to obtain the regional road network within the preset height range based on the road trunk centerlines and intersection polygons.

[0111] In some embodiments, the regional road network includes the traffic attribute information of the road trunk, and the combination unit is further used to determine the traffic attribute information of the road trunk based on the trajectory direction and direction indication information of the crowdsourced trajectory in the regional image, and the trajectory direction of the crowdsourced trajectory is determined based on the collection time corresponding to each trajectory point in the crowdsourced trajectory; based on the center line of the road trunk, the intersection polygon, and the traffic attribute information of the road trunk, the regional road network of the preset height range is obtained.

[0112] In some embodiments, the regional road network includes multiple trunks, and the road network construction device 200 also includes a storage module, which is used to determine the predecessor and successor relationship of each trunk based on the intersection polygons of the regional road network and the traffic attribute information of each trunk connected to the intersection polygon, and the predecessor and successor relationship is used to characterize the connection status of the trunk and other trunks; the predecessor and successor relationship of the trunk is associated with the trunk and stored.

[0113] In some embodiments, the layering module 220 is further used to determine the preset height range to which each trajectory point belongs based on the height of each trajectory point in the crowdsourced trajectory information; and to connect multiple trajectory points within each preset height range based on the collection time corresponding to each trajectory point to obtain the crowdsourced trajectory corresponding to each preset height range.

[0114] In some embodiments, multiple preset height ranges are obtained based on multiple preset height values, and the output module 250 is also used to determine the logical connectivity relationship between the regional road networks based on the preset height range corresponding to each regional road network and the preset height value corresponding to each preset height range; and connect the regional road networks of multiple preset height ranges based on the logical connectivity relationship to obtain the target road network.

[0115] In some embodiments, the conversion module 230 is also used to project each track point in the crowdsourcing track corresponding to the preset height range to a preset grid; based on the position of each track point in the crowdsourcing track corresponding to the preset height range, the environmental perception information corresponding to each track point is projected to the preset grid to obtain an area image corresponding to the preset height range.

[0116] In some implementations, based on the road network construction method provided in the above embodiments, the embodiments of the present application also provide an electronic device.

[0117] As shown in Figure 9, a schematic diagram of the structure of an electronic device provided in an embodiment of the present application is provided. The electronic device 300 includes one or more processors 310; a memory 320; and one or more programs, wherein the one or more programs are stored in the memory 320 and configured to be executed by the one or more processors 310, and the one or more programs are configured to perform the above-described method.

[0118] The electronic device 300 may be a terminal device, and the terminal device may be a computer, a tablet computer, a vehicle-mounted terminal, etc.

[0119] The processor 310 may include one or more processing cores. The processor 310 utilizes various interfaces and circuits to connect the various components within the wearable device. It executes instructions, programs, code sets, or instruction sets stored in the memory 320, as well as accesses data stored in the memory 320, to perform various functions of the wearable device and process data. Optionally, the processor 310 may be implemented using at least one of the following hardware forms: a digital signal processing (DSP), a field-programmable gate array (FPGA), or a programmable logic array (PLA). The processor 310 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing displayed content; and the modem handles wireless communications. It is understood that the modem may not be integrated into the processor 310 and may be implemented separately via a communication chip.

[0120] The memory 320 may include a random access memory (RAM) or a read-only memory (ROM). The memory 320 may be used to store instructions, programs, codes, code sets, or instruction sets. The memory 320 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for implementing at least one function (e.g., a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the following various method embodiments, etc. The data storage area may also store data created by the electronic device 300 during use.

[0121] In some embodiments, the present application further provides a computer-readable storage medium, which stores program code, and the program code can be called by a processor to execute the above method.

[0122] The computer-readable storage medium may be an electronic memory such as a flash memory, an EEPROM (Electrically Erasable Programmable Read-Only Memory), an EPROM, a hard disk, or a ROM. Alternatively, the computer-readable storage medium includes a non-transitory computer-readable storage medium. The computer-readable storage medium has storage space for program code for executing any of the method steps described above. These program codes can be read from or written to one or more computer program products. The program code can be compressed in an appropriate form.

[0123] In the embodiments of the present application, the term "module" or "unit" refers to a computer program or a part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal. It can be implemented in whole or in part by using software, hardware (such as processing circuits or memories), or a combination thereof. Similarly, a processor (or multiple processors or memories) can be used to implement one or more modules or units. In addition, each module or unit can be part of an overall module or unit of the module or unit function.

[0124] The embodiments of the present application have been described above. The above description is illustrative and not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, their practical applications, or improvements to the technology in the market, or to enable other persons skilled in the art to understand the embodiments disclosed herein.

Claims

1. A road network construction method, characterized in that, Including: Obtain the crowdsourcing trajectory information and environmental perception information uploaded by the target device. The crowdsourcing trajectory information includes trajectory point information at multiple moments. Each piece of trajectory point information includes the position and height of the trajectory point. The environmental perception information includes the environmental information collected by the target device corresponding to each trajectory point. Perform hierarchical processing on the crowdsourcing trajectory information based on multiple preset height ranges to obtain the crowdsourcing trajectories corresponding to each preset height range. For each preset height range, based on the crowdsourcing trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourcing trajectory, construct the regional image corresponding to the preset height range. Perform image recognition on the regional image corresponding to the preset height range to obtain the regional road network of the preset height range. Connect the regional road networks of multiple preset height ranges to obtain the target road network.

2. The method according to claim 1, characterized in that, The performing image recognition on the regional image within the preset height range to obtain the regional road network of the preset height range includes: Perform instance segmentation on the regional image within the preset height range to obtain road trunk image blocks and intersection image blocks. Extract the center line based on the road trunk image blocks to obtain the road trunk center line. Perform polygon fitting based on the intersection image blocks to obtain the intersection polygon. Based on the road trunk center line and the intersection polygon, obtain the regional road network of the preset height range.

3. The method according to claim 2, characterized in that, The regional road network includes the traffic attribute information of the road trunk. The obtaining the regional road network of the preset height range based on the road trunk center line and the intersection polygon includes: Determine the traffic attribute information of the road trunk based on the trajectory direction of the crowdsourcing trajectory in the regional image and the direction indication information. The trajectory direction of the crowdsourcing trajectory is determined based on the acquisition time corresponding to each trajectory point in the crowdsourcing trajectory. Based on the road trunk center line, the intersection polygon, and the traffic attribute information of the road trunk, obtain the regional road network of the preset height range.

4. The method according to claim 3, wherein There are multiple road trunks included in the regional road network. After obtaining the regional road network of the preset height range based on the road trunk center line, the intersection polygon, and the traffic attribute information of the road trunk, it further includes: Determine the predecessor-successor relationship of each road trunk based on the intersection polygon of the regional road network and the traffic attribute information of each road trunk connected to the intersection polygon. The predecessor-successor relationship is used to represent the connection situation of the road trunk with other road trunks. Associate and store the predecessor-successor relationship of the road trunk with the road trunk.

5. The method according to claim 1, wherein The performing hierarchical processing on the crowdsourcing trajectory information based on multiple preset height ranges to obtain the crowdsourcing trajectories corresponding to each preset height range includes: Determine the preset height range to which each trajectory point belongs based on the height of each trajectory point in the crowdsourcing trajectory information. Connect the multiple trajectory points within each preset height range based on the acquisition time corresponding to each trajectory point to obtain the crowdsourcing trajectory corresponding to each preset height range.

6. The method according to claim 1, characterized in that The multiple preset height ranges are divided based on multiple preset height values. The connecting the regional road networks of multiple preset height ranges to obtain the target road network includes: Determine the logical connectivity relationship between the regional road networks based on the preset height range corresponding to each regional road network and the preset height value corresponding to each preset height range; Connect the regional road networks of multiple preset height ranges based on the logical connectivity relationship to obtain a target road network.

7. The method according to claim 1, characterized in that The constructing the regional image corresponding to the preset height range based on the crowdsourcing trajectory corresponding to the preset height range and the environmental perception information corresponding to each trajectory point in the crowdsourcing trajectory includes: Project each trajectory point in the crowdsourcing trajectory corresponding to the preset height range onto a preset grid; Based on the positions of the trajectory points in the crowdsourcing trajectory corresponding to the preset height range, project the environmental perception information corresponding to each trajectory point onto the preset grid to obtain the regional image corresponding to the preset height range.

8. A road network construction device, characterized in that, It includes: An acquisition module for acquiring crowdsourcing trajectory information and environmental perception information uploaded by a target device, The crowdsourcing trajectory information includes trajectory point information at multiple moments, each trajectory point information includes the position and height of the trajectory point, and the environmental perception information includes the environmental information collected by the target device at each trajectory point; A layering module for performing layering processing on the crowdsourcing trajectory information based on multiple preset height ranges to obtain the crowdsourcing trajectories corresponding to each preset height range; A conversion module for, for each preset height range, constructing the regional image corresponding to the preset height range based on the crowdsourcing trajectory corresponding to the preset height range and the environmental perception information corresponding to multiple trajectory points in the crowdsourcing trajectory; A processing module for performing image recognition on the regional image corresponding to the preset height range to obtain the regional road network corresponding to the preset height range; An output module for connecting the regional road networks of multiple preset height ranges to obtain a target road network.

9. The device according to claim 8, characterized in that The processing module further includes a segmentation unit, a post-processing unit, and a combination unit; The segmentation unit is used for performing instance segmentation on the regional image within the preset height range to obtain road trunk image blocks and intersection image blocks; The post-processing unit is used for extracting the centerline based on the road trunk image blocks to obtain the road trunk centerline; And performing polygon fitting based on the intersection image blocks to obtain intersection polygons; The combination unit is used for obtaining the regional road network corresponding to the preset height range based on the road trunk centerline and the intersection polygons.

10. The device according to claim 9, wherein: The regional road network includes the traffic attribute information of the road trunk, and the combination unit is further used for determining the traffic attribute information of the road trunk based on the trajectory direction of the crowdsourcing trajectory in the regional image and the direction indication information, and the trajectory direction of the crowdsourcing trajectory is determined based on the acquisition time corresponding to each trajectory point in the crowdsourcing trajectory; obtaining the regional road network corresponding to the preset height range based on the road trunk centerline, the intersection polygons, and the traffic attribute information of the road trunk.

11. The device according to claim 9, characterized in that, There are multiple road trunks included in the regional road network, and the road network construction device further includes: A storage module, and the storage module is used for determining the predecessor-successor relationship of each road trunk based on the intersection polygons of the regional road network and the traffic attribute information of each road trunk connected by the intersection polygons, predecessor-successor The relationship is used to characterize the connection of this road trunk with other road trunks; the predecessor-successor relationship of the road trunk is stored in association with the road trunk.

12. The apparatus according to claim 8, wherein: The layering module is further configured to determine a preset height range to which each trajectory point belongs based on the height of each trajectory point in the crowdsourcing trajectory information; connect a plurality of trajectory points within each preset height range based on the acquisition time corresponding to each trajectory point, so as to obtain a crowdsourcing trajectory corresponding to each preset height range.

13. The apparatus according to claim 8, wherein: The plurality of preset height ranges are obtained by dividing based on a plurality of preset height values, and the output module is further configured to determine the logical connectivity relationship between each regional road network based on the preset height range corresponding to each regional road network and the preset height values corresponding to each preset height range; Connect the regional road networks of a plurality of preset height ranges based on the logical connectivity relationship to obtain a target road network.

14. An electronic device, characterized in that, Comprising: One or more processors; A memory; One or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs are configured to execute the method according to any one of claims 1-7.

15. A computer-readable storage medium, characterized in that, Comprising: The computer-readable storage medium stores program code, and the program code can be called by a processor to execute the method according to any one of claims 1-7.

Images