Intersection functional surface generation method, device and computer readable storage medium

Abstract

The present application provides a method, device and computer-readable storage medium for generating a functional surface of an intersection. Based on the topological relationship of each road segment of a target intersection, the outer boundary lines of adjacent road segments of the target intersection are determined; according to the connection relationship of each of the outer boundary lines, a connection order of multiple clusters reflecting the target intersection is determined, and the multiple clusters are obtained by clustering the features of the road segments of the target intersection; according to the connection order of the multiple clusters, the stop line of the target intersection is connected to the outer boundary line to generate an intersection functional surface corresponding to the target intersection, thereby improving the accuracy of generating the intersection functional surface.

Description

Technical Field

[0001] The present application relates to the field of map technology, and in particular to a method and device for generating a functional surface of an intersection, and a computer-readable storage medium. Background Art

[0002] As autonomous driving technology develops rapidly, accurate identification of intersections has become a crucial part of the autonomous driving system. As a key node in road traffic, accurate identification of intersections can provide key data support for autonomous vehicles' decision-making at intersections, ensuring safe and efficient driving of vehicles and avoiding potential dangers caused by vehicles driving into intersections.

[0003] At present, in the relevant technical field, a more common intersection functional surface generation technology is based on image recognition. This method mainly obtains road image information through image acquisition devices such as vehicle-mounted cameras, and then uses image processing algorithms to identify road boundaries in the image. The identified road boundary information will be further processed to generate an intersection surface model.

[0004] However, this intersection functional surface generation technology based on image recognition has some obvious limitations. Since the image acquisition process is easily affected by the shooting angle, the road boundary recognition is inaccurate. In addition, the obstruction of obstacles on the road is also a problem that cannot be ignored. For example, other vehicles, buildings, trees and other obstructions may partially or completely obstruct the road boundary, making it difficult for the image recognition algorithm to accurately capture the complete boundary information. Another example is the missing road boundary line. These factors greatly reduce the accuracy of the generated intersection surface data, which in turn affects the decision-making reliability of the autonomous driving system at the intersection, bringing potential risks to the safety of autonomous driving.

[0005] Therefore, how to overcome these problems and improve the accuracy of intersection functional surface generation has become a key technical problem that needs to be urgently solved in the development of autonomous driving technology. Summary of the invention

[0006] In order to overcome the problems existing in the related art, this specification provides a method, device and computer-readable storage medium for generating a functional surface of an intersection.

[0007] According to a first aspect of an embodiment of this specification, a method for generating a functional surface of an intersection is provided, the method comprising:

[0008] Based on the topological relationship of each road section of the target intersection, determining the outer edge boundary line of the adjacent road section of the target intersection;

[0009] Determining a connection order of a plurality of clusters of the target intersection according to a connection relationship of each of the outer edge boundary lines, wherein the plurality of clusters are obtained by clustering features of a road section of the target intersection;

[0010] According to the connection sequence of the plurality of clusters, the stop line of the target intersection is connected with the outer boundary line to generate an intersection function surface corresponding to the target intersection.

[0011] According to a method for generating a functional surface of an intersection provided by the present application, the outer edge boundary line of the adjacent road section of the target intersection is determined based on the topological relationship of each road section of the target intersection, including:

[0012] Determining a road surface type of the target intersection based on a topological relationship of the target intersection;

[0013] According to the road surface type, an outer boundary line from an entrance to an exit of an adjacent road section in the target intersection is determined.

[0014] According to a method for generating a functional surface of an intersection provided by the present application, the outer boundary line includes a boundary line corresponding to a turning trajectory and / or a straight trajectory of the vehicle.

[0015] According to a method for generating a functional surface of an intersection provided by the present application, determining the outer edge boundary line from the entrance to the exit of the adjacent road section in the target intersection according to the road surface type includes:

[0016] If the road surface type is a crossroad, determining the right turning trajectory from the entrance to the intersection to the exit of each section of the target intersection as the outer edge boundary line;

[0017] If the road surface type is a T-shaped intersection, determine the right turning trajectory from the entrance to the exit of each section of the target intersection and the straight track from the entrance to the exit of the section where the right turning trajectory is empty as the outer edge boundary line;

[0018] If the road surface type is a straight intersection, it is determined that the straight trajectory corresponding to the empty section of the right turn trajectory of the target intersection from the entrance to the exit is the outer boundary line.

[0019] According to a method for generating a functional surface of an intersection provided by the present application, determining the outer edge boundary line from the entrance to the exit of the adjacent road section in the target intersection according to the road surface type includes:

[0020] Performing perception vectorization processing on the target intersection to obtain lane information of the current road section;

[0021] Determine the rightmost lane in the road section entering the intersection according to the lane information;

[0022] The outer edge dividing line of the right turn and / or straight driving of the rightmost lane is determined according to the road surface type as the outer edge boundary line.

[0023] According to a method for generating a functional surface of an intersection provided by the present application, after performing perception vectorization processing on the target intersection, the method further includes:

[0024] The acquired vectorized road section is cut off using a stop line or a crosswalk to obtain lane information of the current road section after the cut off.

[0025] According to a method for generating a functional surface of an intersection provided by the present application, before determining the connection order of the plurality of clusters of the target intersection according to the connection relationship of each of the outer boundary lines, the method further includes:

[0026] Constructing a virtual area including the center point of the target intersection;

[0027] Clustering is performed on the intersections of the road sections in the target intersection that intersect with the virtual area to generate the plurality of clusters.

[0028] According to a method for generating a functional surface of an intersection provided by the present application, after clustering the intersections of the virtual area and the road sections related to the target intersection to generate the multiple clusters, the method further includes:

[0029] In response to the number of the plurality of clusters being less than a threshold, the clusters associated with the target intersection are supplemented according to a symmetry principle so that the number of the plurality of clusters is greater than or equal to the threshold.

[0030] According to a method for generating a functional surface of an intersection provided by the present application, the step of determining the outer boundary line of an adjacent road segment of the target intersection includes:

[0031] The following processing is performed on each of the multiple clusters:

[0032] Determining the type of the cluster according to the identifier of the road segment corresponding to the cluster;

[0033] In response to the type of the cluster being a supplementary cluster, determining a boundary line of a straight track from an entrance to an exit from an intersection corresponding to the supplementary cluster as the outer boundary line;

[0034] In response to the type of the cluster being a real cluster, a boundary line of a turning trajectory from an entrance to an exit from an intersection corresponding to the real cluster is determined as the outer boundary line.

[0035] The present application also provides a device for generating a functional surface of an intersection, the device comprising:

[0036] A first information processing module, configured to determine an outer edge boundary line of an adjacent road segment of the target intersection based on a topological relationship between road segments of the target intersection;

[0037] A second information processing module is used to determine the connection order of multiple clusters reflecting the overall structure and traffic flow direction of the target intersection according to the connection relationship of each of the outer boundary lines;

[0038] The intersection functional surface generating module is used to connect the stop line of the target intersection with the outer boundary line according to the connection order of the multiple clusters, so as to generate the intersection functional surface corresponding to the target intersection.

[0039] The present application also provides a device for generating a functional surface of an intersection, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for generating a functional surface of an intersection as described above is implemented.

[0040] The present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method for generating a functional surface of an intersection as described in any one of the above is implemented.

[0041] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements the method for generating a functional surface of an intersection as described in any one of the above.

[0042] Compared with the inaccurate intersection surface data caused by the current generation of intersection surfaces by perception, the method, device and computer-readable storage medium for generating intersection functional surfaces in the embodiments of this specification have the following beneficial effects:

[0043] On the first aspect, based on the topological relationship of each road section of the target intersection, the outer boundary line of the adjacent road section of the target intersection is determined, wherein the topological relationship is established based on the actual connection status of the road sections, and the overall structure of the target intersection and the connection relationship between each road section are fully considered. The outer boundary line determined by the topological relationship can better reflect the actual traffic situation of the target intersection, so that the subsequently generated intersection functional surface is more in line with the actual traffic needs, thereby improving the accuracy of the intersection functional surface generation.

[0044] In the second aspect, the connection order of the multiple clusters of the target intersection is determined according to the connection relationship of each of the outer boundary lines, and the multiple clusters are clustered by the characteristics of the road section of the target intersection, reflecting the driving needs of vehicles in different directions at different target intersections. Subsequently, according to the connection order of the multiple clusters, the stop line of the target intersection is connected to the outer boundary line to ensure that the generated intersection function surface corresponding to the target intersection meets the traffic rules and actual traffic needs. In addition, the generation process has logical coherence, avoiding traffic chaos, and further improving the accuracy of the intersection function surface.

[0045] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the specification and, together with the description, serve to explain the principles of the specification.

[0047] Figure 1 is a flowchart of a method for generating a functional surface of an intersection according to an exemplary embodiment of the present specification;

[0048] Figure 2 is a topological relationship of a target intersection shown in this specification according to an exemplary embodiment;

[0049] Figure 3 is a topological relationship of a target intersection including detailed road section information shown in this specification according to an exemplary embodiment;

[0050] Figure 4 is a schematic diagram of a functional surface of an intersection according to an exemplary embodiment of this specification;

[0051] Figure 5 is a schematic diagram of a device for generating a functional surface of an intersection according to an exemplary embodiment of the present specification;

[0052] Figure 6 It is a schematic block diagram of a device for generating a functional surface of an intersection according to an exemplary embodiment of the present specification. DETAILED DESCRIPTION

[0053] Here, the technical solutions in the embodiments (or "implementation methods") of the present application will be described clearly and completely in conjunction with the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements.

[0054] If there are terms involving directional indications or positional relationships in the embodiments of the present application (such as up, down, left, right, front, back, inside, outside, top, bottom, center, vertical, horizontal, longitudinal, transverse, length, width, counterclockwise, clockwise, axial, radial, circumferential, etc.), such terms are only used to explain the relative positional relationship, movement, etc. between the components under a certain specific posture (as shown in the accompanying drawings); if the specific posture changes, the directional indication or positional relationship will also change accordingly. In addition, the terms "first" and "second" in the embodiments of the present application are only used for the purpose of convenience of description and cannot be understood as indicating or implying relative importance.

[0055] The present application provides a method, device and computer readable storage medium for generating a functional surface of an intersection. The present application is described in detail below in conjunction with the accompanying drawings. The features of the following embodiments and implementations can be combined with each other.

[0056] In the intersection surface perception stage, various sensors and data collection methods are usually used to obtain basic information about the intersection, such as road shape, lane distribution, traffic signs, etc. The generation of the intersection functional surface is based on this, further considering the actual traffic functions and needs of the intersection, and making more detailed divisions and annotations of the intersection surface.

[0057] In current technology, the functional surface of the intersection is sensed by on-board equipment (such as cameras, lidar, etc.), and the intelligent driving vehicle performs perception simultaneously during its driving process. However, due to the low perception efficiency and the possibility of being affected by angle deviation, occlusion, etc. during the perception process, for example, other vehicles, buildings, trees and other obstructions may partially or completely block the road boundary, making it difficult for the image recognition algorithm to accurately capture the complete boundary information, which in turn leads to low accuracy in the generated intersection surface data, affecting the reliability of autonomous driving decisions at intersections and posing potential risks to the safety of autonomous driving.

[0058] In order to solve the above technical problems, this specification provides a method for generating a functional surface of an intersection.

[0059] By sensing the stop lines and physical separations of the road sections after vectorization, the functional surface of the intersection is constructed and output.

[0060] Specifically, the outer boundary line of the target intersection is determined through topological relationships, and the outer boundary line is connected with the vectorized stop line and outer boundary line of the target intersection according to a certain order to generate an intersection functional surface, rather than obtaining the intersection functional surface through a perception process, thereby improving the accuracy and completeness of the intersection surface data.

[0061] The intersection functional surface generation method in this article generates the intersection functional surface of the intersection ahead based on the prior information obtained, which is used as a backup information for intelligent driving vehicle perception to improve the accuracy of the intersection surface data.

[0062] It should be noted that the intersection functional surface generation solution in this article is applicable to any device or system in terminals, vehicles, independent servers, server clusters, cloud, vehicle network platforms, Internet of Things platforms, etc.

[0063] The present application provides an embodiment of a method for generating a functional surface of an intersection, referring to Figure 1 , Figure 1 : is a flow chart of a method for generating a functional surface of an intersection provided in an embodiment of this specification, comprising the following steps:

[0064] S100, determining outer edge boundary lines of adjacent road segments of the target intersection based on the topological relationship between the road segments of the target intersection.

[0065] In a traffic system, a target intersection may be any intersection in a traffic network, such as an intersection in a traffic network having traffic flow and a specific road connection relationship, wherein traffic flow refers to the number of vehicles, pedestrians or other traffic participants passing through an intersection within a certain period of time, and a specific road connection relationship refers to the connection method and layout between the intersection and surrounding roads. It can be understood that a target intersection may be a location where at least two roads intersect or connect or an area containing the location, and the area where the target intersection is located may have any one or a combination of knowledge information such as road markings, traffic signs, boundary signs, etc. For example, in a traffic network, a crossroads, a straight intersection, a T-intersection, and a three-way intersection may all be target intersections in embodiments of the present application.

[0066] In some embodiments, the target intersection in the present application may include different types of intersections, such as the target intersection may include any one or more of a straight intersection, a T-shaped intersection, a three-way intersection, a crossroads, and other complex intersections.

[0067] The topological relationship in this article is a graphical structure that represents the connection relationship between various elements in the road network. In the field of transportation, topological relationships are usually used to describe the connection mode and relative position relationship between elements such as intersections, road sections, and traffic signs. Through the obtained topological relationship, the adjacent sections of the target intersection can be determined. The adjacent sections in the embodiment of the present application can be sections that enter or leave the target intersection from different directions. The outer edge boundary lines of the adjacent sections are further determined for the subsequent generation of the intersection functional surface.

[0068] The outer boundary line described herein refers to the boundary line of the connecting part of the adjacent road section and the target intersection. The outer boundary line can be used to distinguish the range of different road sections and determine the actual boundary of the intersection. It can be understood that the outer boundary line in the embodiment of the present application can reflect any one or more obstacles such as green belts, curbs or other objects at the edge of the target intersection. During the intelligent driving process, when the vehicle enters the functional surface of the intersection, its driving trajectory cannot exceed the outer boundary line.

[0069] In some embodiments, the outer boundary line includes a boundary line corresponding to the turning trajectory of the vehicle, such as the right turn boundary line of the right turn trajectory in the intersection, and the outer boundary line also includes a boundary line corresponding to the straight track, such as the right side boundary line of the straight track in the straight intersection. It can be understood that the above-mentioned right turn or right side boundary line as the outer boundary line of the target intersection is set according to the traffic rules of driving on the right. In other examples, the left turn boundary line of the left turn trajectory and / or the left side boundary line of the straight track can also be set as the actual boundary of the target intersection according to the traffic rules of driving on the left. No specific limitation is made here. The following is explained by taking driving on the right as an example. The specific implementation methods of other forms are basically the same and will not be repeated here.

[0070] In this embodiment, the outer boundary line is determined based on the topological relationship of the target intersection obtained from prior information. Compared with the traditional method of using perception, the topological relationship has higher accuracy, and the corresponding outer boundary line is not affected by perceptual deviations, obstacle occlusions, etc. during perception, thereby improving the accuracy and reliability of the outer boundary line in the generated intersection functional surface.

[0071] S200, determining a connection order of a plurality of clusters of the target intersection according to a connection relationship of each of the outer boundary lines, wherein the plurality of clusters are obtained by clustering features of a road section of the target intersection;

[0072] The cluster described in this article refers to a set of road segments with similar features obtained by clustering the features of each road segment of the target intersection. These similar features may include the geometry of the road segment, traffic function, traffic flow, road grade, etc. The clustering process of each road segment will be described in detail later.

[0073] By clustering road sections into different clusters, the complex structure of the target intersection can be clearly understood, and the functions and relationships of the various parts of the target intersection can be intuitively seen to reflect the overall structure and traffic flow of the target intersection. The connection sequence of the outer boundary line and the stop line used to determine the generation process of the intersection functional surface can ensure that the generated intersection functional surface is consistent with the actual traffic flow direction, making the layout of the target intersection more reasonable.

[0074] S300, connecting the stop line of the target intersection with the outer boundary line according to the connection order of the plurality of clusters, to generate an intersection function surface corresponding to the target intersection.

[0075] The stop line mentioned in this article refers to the location at the target intersection, before the crosswalk, or other locations where vehicles need to stop. It is used to indicate that vehicles must stop at a specific location to wait for traffic signals, avoid pedestrians or other vehicles with priority, and prevent vehicles from entering the target intersection in disorder and causing traffic chaos. For example, at an intersection controlled by a traffic light, vehicles must wait for the green light signal behind the stop line before passing. The determination of the stop line will be explained in detail in the subsequent chapters.

[0076] After the connection sequence of the stop lines, outer boundary lines and clusters is determined, adjacent stop lines and outer boundary lines are connected based on the connection sequence, thereby generating an intersection functional surface corresponding to the target intersection.

[0077] In this embodiment, the purpose is to generate the intersection functional surface by vectorizing the outer edge boundary line and the stop line, so as to improve the accuracy of the intersection surface generation.

[0078] This application is mainly used in the perception of intersection surfaces by intelligent driving vehicles, from the complete process of collecting data from the target intersection to generating the intersection functional surface, and working in coordination with the intersection surface perceived by intelligent driving. It can be understood that when the intelligent driving system currently on the market perceives the intersection surface, the intersection functional surface generated based on this application can be used to verify the accuracy of the intersection surface generated by perception, and improve the reliability and safety of intelligent driving.

[0079] The intersection functional surface generation method involves determining the stop line of the target intersection, determining the outer boundary line of the target intersection, connecting the stop line and the outer boundary line according to the order of each cluster of the target intersection, and generating the intersection functional surface. These three parts are closely combined to form a set of efficient, accurate and complete intersection functional surface generation solutions, which is of great significance for improving the safety performance of intelligent driving vehicles.

[0080] In some embodiments, based on the topological relationship of each road segment of the target intersection, the process of determining the outer edge boundary line of the adjacent road segment of the target intersection is as follows:

[0081] S101, determining the topological relationship of each road section at the target intersection.

[0082] In step S101, the topological relationship may be determined according to the road line type of the target intersection.

[0083] Specifically, refer to Figure 2 , Figure 2 It is the topological relationship of the target intersection. The road line type usually refers to the geometric shape of the road segment in space. The road segment can be a straight line, a curve, or a complex shape composed of multiple line segments.

[0084] By analyzing the relationship between the linear directions of each road segment around the target intersection in the road linear type and the target intersection, the topological relationship of the target intersection is gradually constructed, wherein the linear direction is the direction in which the road segment extends.

[0085] As an example, the specific process of determining the topological relationship by the road line direction is as follows:

[0086] The location of the target intersection and the road section information within the area where the target intersection is located are obtained, including any one or any combination of the coordinate position, linear shape, etc. of the road section.

[0087] Determine the road line direction, and analyze the line direction of each road section: determine the direction of the local road section by calculating the direction of the line connecting adjacent points on the road section;

[0088] Determine the connection relationship between the target intersection and the road linear direction, check the road linear direction around the target intersection, and determine the connection relationship between the road section corresponding to the road linear direction and the target intersection in the direction by whether the road linear direction is consistent with a certain direction of the intersection;

[0089] Determine the relationship between adjacent road sections, analyze the linear directions of adjacent road sections, and determine the connection method between them;

[0090] Constructing a topological relationship: According to the connection relationship between the target intersection and the road section and the relationship between adjacent road sections, constructing a topological relationship of the target intersection, wherein the node represents the intersection, the edge represents the road section, and the direction of the edge represents the traffic direction of the road section.

[0091] In some other embodiments, in step S101, the topological relationship of each road section of the target intersection may also be obtained from a reference map (the reference map may include but is not limited to a standard definition, standard precision or high precision map, etc.).

[0092] It should be understood that the topological relationship in the standard map / high-precision map is also constructed by considering the connection judgment between the target intersection and the road section and the relationship between adjacent road sections; because the topological relationship obtained by the standard map / high-precision map is usually professionally verified and other factors such as traffic signs are considered, it has higher accuracy and reliability.

[0093] Therefore, as an embodiment, in the embodiment of the present application, the topological relationship of each road section of the target intersection can also be preferably obtained from the standard map / high-definition map. When there is doubt about the road section in a specific area, or the standard map / high-definition map is not updated in time or the data is not detailed enough, the topological relationship can be determined by the road line type. Both methods of obtaining the topological relationship are acceptable and can be selected according to actual needs without specific limitation.

[0094] S102: Determine the outer boundary line of the target intersection according to the topological relationship.

[0095] The above topological relationship can show the connection status of each section of the target intersection. By analyzing the topological relationship, it is clear which sections are dedicated right-turn sections or sections with right-turn function at the target intersection, which sections are dedicated straight-through sections, etc. By analyzing the topological relationship of the target intersection and the vector processing of the vehicle's perception of the target intersection, the following is obtained: Figure 3 The topological relationship of the detailed road section information shown, the updated topological relationship includes three levels of road information of road section, lane line, and lane boundary line devide, which are used to determine the outer boundary line of the target intersection.

[0096] In some embodiments, the step of determining the outer boundary line of the target intersection according to the topological relationship includes the following steps:

[0097] S1021: Determine a road surface type of the target intersection based on a topological relationship of the target intersection.

[0098] In S1021, the type of the target intersection determined according to the topological relationship includes a straight intersection, a T-shaped intersection, a three-way intersection, a crossroad, and any other complex intersections. According to the road surface type of the target intersection, it is determined which boundary lines in the target intersection are used for subsequent generation of the intersection functional surface.

[0099] S1022: Determine, according to the road surface type, an outer boundary line from an entrance to an exit of an adjacent road section in the target intersection.

[0100] In S1022, as described above, the outer boundary line includes the boundary line corresponding to the turning trajectory of the vehicle and the boundary line corresponding to the straight trajectory. The turning trajectory includes the left turning trajectory and the right turning trajectory. The functional surface of the intersection in this article is within the scope of the intersection, and the boundary of the intersection is circled. Figure 3 Although the information such as the boundary line of the left-turn lane of the vehicle is also displayed, the trajectory of the vehicle when turning left is within the intersection surface and does not belong to the edge position of the intersection surface. Therefore, the boundary line of the left-turn trajectory is not considered for the time being. Therefore, when determining the outer boundary line below, the boundary line of the right-turn trajectory and the boundary line of the straight-ahead trajectory are mainly determined.

[0101] Specifically, determining the outer edge boundary line from the entrance to the exit of the adjacent road section in the target intersection according to the road surface type includes the following steps:

[0102] Step a1, performing perception vectorization processing on the target intersection to obtain lane information of the current road section.

[0103] The topological relationship of the target intersection includes basic information at the road level, showing the general direction of the road, the location of the intersection, the level of different roads, etc. The vectorized data provides more detailed information about the target intersection and road. Through the vehicle's sensors and data processing technology, upstream perception vectorization can obtain specific lane information including the precise shape of the road, lane division, lane boundary lines, road dividing lines, stop lines, etc. Upstream perception vectorization can quickly reflect the current status of the road by collecting data in real time.

[0104] Step a2: determining the rightmost lane in the road section entering the intersection according to the lane information.

[0105] When the vehicle is turning right or going straight, the boundary line of the target intersection is determined based on the boundary line of the lane the vehicle is traveling in. Usually, under the traffic rule of driving on the right, the outer boundary line of the vehicle is based on the boundary line of the rightmost lane. Therefore, the lane information obtained by upstream perception vectorization is used to determine the rightmost lane of the vehicle when it is traveling on the current road section, including the rightmost lane corresponding to the vehicle entering and leaving the target intersection.

[0106] For multiple lane boundary lines, an outer edge boundary line for generating an intersection functional surface of the target intersection is determined according to the driving trajectory of the target intersection.

[0107] As an example, during a right turn, after finding the information of the right line right-devide of the rightmost lane of the current section through road, lane line, and boundary line devide layer by layer, the line between the right line and the right line of the lane leaving the target intersection in the section after the right turn is the outer boundary line of the target intersection on the turning trajectory.

[0108] During the straight-ahead process, after determining the right line of the rightmost straight-ahead lane in the current section through layer-by-layer information search, the line connecting the right line and the right line of the lane leaving the target intersection in the section after passing the target intersection is the outer edge boundary line of the target intersection on the straight-ahead trajectory.

[0109] Therefore, when selecting the right-turn trajectory or the straight-ahead trajectory of the vehicle to determine the outer edge boundary line, it needs to be determined according to the actual road surface type of the target intersection.

[0110] Step a3: determining the outer edge dividing line of the right turn and / or straight driving of the rightmost lane as the outer edge boundary line according to the road surface type.

[0111] Depending on the type of road surface, the outer edge boundary line determined includes the following situations:

[0112] Case 1: If the road type is a crossroad, the right turning trajectory from the entrance to the exit of each section of the target intersection is determined as the outer edge boundary line. Figure 3 As shown, the line connecting the right line of the rightmost right-turn lane of each road section and the right line of the lane leaving the target intersection in the road section after the right turn is the outer edge boundary line.

[0113] Case 2: If the road surface type is a T-shaped intersection, the right turn trajectory from the entrance to the exit of each section of the target intersection and the straight track corresponding to the entrance to the exit of the section where the right turn trajectory is empty are determined as the outer boundary line. Among them, the straight track corresponding to the entrance to the exit of the section where the right turn trajectory is empty is the outer boundary line, which means that the line between the right line of the rightmost straight lane in the straight track and the right line of the lane leaving the target intersection in the section after passing the target intersection is the outer boundary line of the target intersection on the straight track.

[0114] Case 3: If the road surface type is a straight-through intersection, the straight-through trajectory corresponding to the section from the entrance to the exit of the target intersection where the right-turn trajectory is empty is determined as the outer boundary line. The target intersection has no turning trajectory, so the outer boundary line is obtained by determining the line connecting the right line of the rightmost straight lane in the straight-through trajectory and the right line of the lane leaving the target intersection in the section after the straight-through passes the target intersection.

[0115] It is worth noting that, except for the three road surface types listed above, the specific implementation methods of other road surface types are basically the same and will not be repeated here.

[0116] In some embodiments, a second-order Bezier curve is used to determine the connection line between the right-turn trajectory or the straight-ahead trajectory to obtain an outer boundary line, thereby ensuring that the outer boundary line after connection is smoother, thereby reducing the bumps and discomfort of the vehicle during driving.

[0117] In this embodiment, the outer boundary line is generated by vectorization, and the vectorized data can provide the intelligent driving system with more detailed road environment perception and help the vehicle make more accurate decisions.

[0118] In other embodiments, since there may be road sections extending into the target intersection in the initially acquired image of the target intersection, it is necessary to optimize each road section to obtain a complete functional surface of the intersection.

[0119] As an example, after performing the perceptual vectorization processing on the target intersection, the method further includes:

[0120] The acquired vectorized road section is cut off using a stop line or a crosswalk to obtain lane information of the current road section after the cut off.

[0121] Among them, the stop line or crosswalk is obtained through upstream perception vectorization processing, and each road section is cut off by the stop line or crosswalk to ensure that the generated intersection function surface conforms to the actual scene and improve its accuracy. In addition, when generating the outer edge boundary line, the intersection point of the stop line and the right turn lane or through lane where the vehicle is located can also be used as a control point to obtain the outer edge boundary line using a second-order Bezier curve.

[0122] It is worth noting that the stop line plays an important role in intersection traffic. It is used to indicate the stop position of vehicles at the intersection to ensure the order and safety of traffic. However, in actual situations, the stop line may be missing, which may be caused by road construction, wear and tear, inaccurate map data, etc. In some embodiments, the stop line is determined as follows:

[0123] The stop line of the target intersection is sensed, and when a complete stop line is sensed, the stop line is vectorized for subsequent generation of the intersection functional surface; when a missing stop line is sensed or the stop line cannot be sensed, the sensed stop line is supplemented.

[0124] Specific supplementary stop line methods include:

[0125] Check whether there is a crosswalk on the road section where the stop line is located. If there is a crosswalk, a virtual stop line is generated at a certain distance from the crosswalk. The two ends of the virtual stop line cannot exceed the boundary lines of the left and right ends of the road section. If there is no crosswalk, check whether there is a real stop line on the road section symmetrical to the current road section. If so, add a virtual stop line according to the principle of symmetry.

[0126] In other examples, a virtual stop line can be generated based on the baseline of the cluster where the current road section is located. The baseline can be the center line of the road in the cluster that reflects the direction and distribution of the road. In the absence of a real stop line, a virtual stop line is generated based on the center line of the road. It can be considered to set the virtual stop line at a certain distance from the center line and perpendicular to the road to simulate the actual traffic stop position. The baseline can also be the edge line of the road section. At a position close to the target intersection, a virtual stop line is set at a certain distance from the edge line of the road section to ensure that the vehicle can have a clear stop position before entering the target intersection. The baseline can also be an average driving trajectory. This trajectory can reflect the usual driving path of the vehicle on the road section. Generating a virtual stop line based on the average driving trajectory can make the stop line more in line with the actual traffic conditions and improve the safety and efficiency of intersection traffic.

[0127] After the vectorized outer boundary line and stop line are determined through the above embodiment, the connection order of the stop line and the outer boundary line in the intersection functional surface is further determined. The connection order can be determined according to the order in which the various road sections in the target intersection are clustered.

[0128] Specifically, in some embodiments, before determining the connection order of multiple clusters reflecting the overall structure and traffic flow direction of the target intersection according to the connection relationship of each of the outer boundary lines in step S200, the method further includes the following steps:

[0129] S401, constructing a virtual area including the center point of the target intersection.

[0130] In S401, the intersection center point of the target intersection is obtained from the standard precision map or the high precision map, and a virtual area is created with the intersection center point and a preset radius. The shape of the virtual area is not limited, such as the shape of the virtual area can be any shape of a circle, a square, a rectangle, a regular polygon, an irregular polygon, etc. For ease of understanding, two specific examples of virtual areas are given here: the virtual area can be a circle constructed with the intersection center point as the center and a preset radius as the radius, and the virtual area can also be a square generated with the intersection center point as the center point and a preset side length as the side length, and the preset side length can be determined based on the aforementioned preset radius, etc. Among them, the preset radius is less than the length of all road sections related to the target intersection, and the virtual area created thereby completely covers the range of the target intersection.

[0131] S402: Clustering the intersections of the road sections intersecting the virtual area at the target intersection to generate the plurality of clusters.

[0132] In S402, clustering is performed on the intersections between the virtual area and the road sections related to the target intersection, and the road sections with similar characteristics are clustered into a cluster, so that the overall structure and traffic flow direction of the target intersection are reflected through the cluster.

[0133] That is, different road sections in the same cluster in the embodiment of the present application have similar characteristics. By clustering the relevant road sections of the target intersection, the cluster in the embodiment of the present application can better express the overall structure of the target intersection, which helps to clarify the position and relationship of each road section in the intersection, thereby providing a basis for generating an accurate intersection functional surface.

[0134] Specifically, the intersection distribution characteristics of multiple intersections where the virtual area intersects with the road section are observed, and the intersection distribution characteristics express the different locations of the target intersection and the potential traffic flow direction. The aforementioned intersection distribution characteristics may include, but are not limited to, the location and number of intersections, which can be used to preliminarily determine possible clustering areas. For example, if the intersection distribution characteristics express that there are more intersections concentrated in one area in a certain direction, then this area may become a candidate for a cluster.

[0135] In some embodiments, clustering is performed using the DBSCAN algorithm, which is a density-based clustering algorithm whose core idea is to cluster closely connected points in space into a cluster. In this intersection clustering scenario, it can determine the clustering situation based on the distribution density of the intersections. If the number of intersections in a certain area reaches a set density threshold, then these intersections will be divided into a cluster.

[0136] The specific process includes the following steps:

[0137] Step b1, determining the distance between the virtual area and each intersection point of all road sections related to the target intersection.

[0138] First, calculate the distance between each intersection point. You can use common methods such as Euclidean distance to determine the distance between two intersection points in space.

[0139] Step b2: Group the road sections corresponding to the intersections whose distances are less than or equal to the distance threshold into the same cluster.

[0140] If the distance between some intersections is less than the set distance threshold, then these intersections are considered to have a high enough density and can be clustered. When there is an additional auxiliary road in a certain direction, an intersection will be added. If the distance between this new intersection and other intersections also meets the density requirement, then these three intersections will be clustered together. This is because the DBSCAN algorithm will classify similar points into the same cluster based on density. Even if a new intersection appears, as long as it is close enough to the existing intersection, it will be included in the same cluster.

[0141] Step b3: generating a plurality of said clusters.

[0142] Through the above clustering process, each road section of the target intersection is clustered into multiple clusters, and the clusters reflect the structure of the target intersection.

[0143] In another embodiment, there may be a situation where the distance between adjacent road sections in the target intersection is relatively close, but the traffic flow directions of the two are different. In order to improve the analysis accuracy of the target intersection, the clustering process needs to be further optimized.

[0144] Specifically, after clustering the road sections corresponding to the intersections whose distances are less than or equal to the threshold into the same cluster, the method further includes:

[0145] Step b21, determining the vector between the center point of the target intersection and the intersection point after preliminary clustering processing.

[0146] A coordinate system is established with the center point of the intersection as the origin. For each road section, the direction angle is determined from the center point to the intersection of the road section and the center point. For example, the angle formed by the intersection point corresponding to a road section in the coordinate system and the center point is 180 degrees, and the angle formed by the intersection point corresponding to another road section and the center point is 150 degrees.

[0147] Step b22, grouping the road sections corresponding to the intersections whose vector angles are less than or equal to the angle threshold into the same cluster.

[0148] If the angles of two road segments differ greatly, it means that they are in different directions. In this case, even if the intersections of the two road segments meet the clustering condition in terms of distance, they cannot be clustered into one cluster due to their different directions. For example, in the previous example, the two road segments differ by 30 degrees. According to the set direction difference threshold, if the threshold is exceeded, they cannot be clustered into the same cluster.

[0149] In this embodiment, in the clustering process, both the distance between the intersections and the direction of the road sections are considered. Only when the distance between the intersections is less than the distance threshold and the direction difference is within an acceptable range, the road sections corresponding to these intersections can be clustered into one cluster. This can more accurately reflect the connection relationship and traffic flow direction of different road sections in the actual intersection, and avoid incorrectly clustering road sections with large direction differences.

[0150] After clustering the intersections of the virtual area and the road sections related to the target intersection to generate the plurality of clusters, the method further includes:

[0151] In response to the number of the plurality of clusters being less than a threshold, the clusters associated with the target intersection are supplemented according to a symmetry principle so that the number of the plurality of clusters is greater than or equal to the threshold.

[0152] The road sections around the intersection are gradually clustered, and finally four relatively independent but interrelated clusters are formed. These four clusters can represent different directions or functional areas of the intersection, respectively, and provide a basic structural framework for the subsequent intersection surface generation. Therefore, the threshold described in this article can be set based on actual needs, and can be but not limited to being set based on any one or more actual demand requirements of the generation quality, accuracy, efficiency, etc. of the intersection functional surface, such as being but not limited to setting the threshold to 3, 4, or 5, etc.

[0153] At intersections, traffic flows in different directions need to be smoothly converted and connected. In special intersections such as T-junctions and straight intersections, the intersection surface is incomplete. Supplementary clusters can fill these possible gaps and enhance the connectivity, making the subsequent generated intersection surface more continuous and complete, so that vehicles and pedestrians can pass through the target intersection more safely and efficiently. In addition, the intersection functional surface generated in this way is more visually accurate and can better reflect the actual road conditions, providing a more reliable foundation for subsequent traffic planning, navigation and other applications. Therefore, the missing clusters are supplemented based on the comparison of the number of multiple clusters with the threshold.

[0154] As an example, if there is no supplementary cluster on the upper side of the T-junction, there may be a gap in the intersection surface. By adding supplementary clusters based on the principle of central symmetry and connecting them with the outer edge of the straight section, this gap can be filled, making the target intersection surface smoother. Then, the connection order of the clusters is analyzed for the complete intersection surface.

[0155] After clustering, the connection mode of each road section in the target intersection is determined according to the direction of the cluster, especially for the determination of the right turn direction, which can clarify the driving path of the vehicle at the target intersection. The stop lines connecting the intersections are to ensure the orderly stop and start of traffic flow at the intersection, and clustering provides a logical basis for this connection. Therefore, after the order of the clusters is determined, the connection order of the stop lines can be established according to this order.

[0156] Specifically, the order among the plurality of clusters is determined according to the connection relationship of the right turns of each road section of the target intersection, and the order includes a clockwise order and a counterclockwise order.

[0157] According to the clockwise or counterclockwise order of the determined clusters, the stop lines are completed to generate a complete intersection functional surface.

[0158] It can be understood that each cluster has an entry and exit stop line, which can be used as the key nodes for connecting clusters. The entry and exit stop lines of adjacent clusters are connected in sequence through the outer boundary line to construct the overall framework of the exit. For example, the exit stop line of the first cluster is connected to the entry stop line of the second cluster through the outer boundary line, and so on, forming a continuous and orderly road network structure, such as Figure 4 shown.

[0159] This can avoid confusion in connections and ensure that the generated road network is logical and coherent. For example, if the order of clusters is counterclockwise, the connection of stop lines should also be carried out in a clockwise order, connecting the stop lines of adjacent clusters in sequence.

[0160] In some embodiments, when special intersections are clustered, clusters will be supplemented when the number of clusters is less than a threshold, and these supplemented clusters are virtual. There are no corresponding road sections and stop lines on the road sections for the virtual clusters. Therefore, when generating the functional surface, it is necessary to combine the actual intersection structure to connect the stop lines corresponding to the real clusters with the outer boundary lines.

[0161] Therefore, in some examples, when clustering the road segments, relevant attribute information is added to each cluster to better describe and analyze the characteristics of the cluster. For example, the type of cluster can be added (such as real clusters, supplementary clusters), and of course, attributes such as traffic flow, number of lanes, and intersection names can also be added.

[0162] Specifically, the step of determining the outer boundary line of the adjacent road segment of the target intersection includes:

[0163] The following processing is performed on each of the multiple clusters:

[0164] Step c1, determining the type of the cluster according to the identifier of the road section corresponding to the cluster;

[0165] In the embodiment of the present application, the clusters obtained by clustering include two types: real clusters and supplementary clusters. Then, in step c1, each cluster can be associated with the ID of the corresponding real road segment. By judging whether the road segment in the cluster has an ID identifier, the type of cluster is determined to be a real cluster or a supplementary cluster. The stop line and outer edge boundary line in the real cluster are used to generate a complete intersection functional surface.

[0166] Step c2, in response to the type of the cluster being a supplementary cluster, determining the boundary line of the straight-line trajectory from the entrance to the exit corresponding to the supplementary cluster as the outer boundary line; or, in response to the type of the cluster being a real cluster, determining the boundary line of the turning trajectory from the entrance to the exit corresponding to the real cluster as the outer boundary line.

[0167] That is, in step c2, for the supplementary cluster among the multiple clusters, the boundary line of the straight-line trajectory from the entrance to the exit intersection corresponding to the cluster of this type is determined as the outer boundary line; for the real cluster among the multiple clusters, the boundary line of the turning trajectory from the entrance to the exit intersection corresponding to the cluster of this type is determined as the outer boundary line.

[0168] There is no stop line and outer edge boundary line in the supplemented cluster, and the boundary line on this side can be determined based on the boundary line of the straight track from the entrance to the intersection to the exit corresponding to the supplemented cluster. In other words, when generating the intersection functional surface, the overall outer edge boundary line is determined based on the boundary line of the right turn track in the real cluster and the boundary line of the straight track corresponding to the supplemented cluster. Then, according to the connection order of the clusters determined above, the outer edge boundary line and the stop line are connected to generate a complete intersection functional surface.

[0169] As an example, for a T-shaped intersection, the upper side is a supplementary cluster, and when generating the intersection surface, the outer edge boundary line of the upper straight section is used for connection.

[0170] Specifically, in the case of a T-junction, there may be a supplementary cluster on the upper side. This supplementary cluster may be added to make the intersection surface more complete. When generating the intersection surface, the outer edge boundary line of the straight section is used to connect this supplementary cluster. This ensures that the intersection surface can also remain continuous and complete at the T-junction.

[0171] The present application provides a method, device and computer-readable storage medium for generating a functional intersection surface, which have the following beneficial effects compared with the current method of generating an intersection surface by perception, which results in insufficient intersection surface data accuracy:

[0172] On the first aspect, based on the topological relationship of each road section of the target intersection, the outer boundary line of the adjacent road section of the target intersection is determined, wherein the topological relationship is established based on the actual connection status of the road sections, and the overall structure of the target intersection and the connection relationship between each road section are fully considered. The outer boundary line determined by the topological relationship can better reflect the actual traffic situation of the target intersection, so that the subsequently generated intersection functional surface is more in line with the actual traffic needs, thereby improving the accuracy of the intersection functional surface generation.

[0173] In the second aspect, the connection order of the multiple clusters of the target intersection is determined according to the connection relationship of each of the outer boundary lines. The multiple clusters are obtained by clustering the characteristics of the road sections of the target intersection, reflecting the driving needs of vehicles in different directions at different target intersections. Subsequently, according to the connection order of the multiple clusters, the stop line of the target intersection is connected to the outer boundary line to ensure that the generated intersection functional surface corresponding to the target intersection meets traffic rules and actual traffic needs. In addition, the generation process has logical coherence, avoids traffic chaos, and further improves the accuracy of the intersection functional surface.

[0174] Based on the same application concept as the above method, the embodiment of the present application also proposes a device for generating a functional surface of an intersection, such as Figure 5 shown.

[0175] The device comprises:

[0176] A first information processing module 502, configured to determine an outer edge boundary line of an adjacent road segment of the target intersection based on a topological relationship between road segments of the target intersection;

[0177] A second information processing module 504 is used to determine the connection order of multiple clusters reflecting the overall structure and traffic flow direction of the target intersection according to the connection relationship of each of the outer boundary lines;

[0178] The intersection functional surface generating module 506 is used to connect the stop line of the target intersection with the outer boundary line according to the connection sequence of the multiple clusters to generate an intersection functional surface corresponding to the target intersection.

[0179] The implementation process of the functions and effects of each module / submodule / unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and the same technical effect can be achieved, so it will not be repeated here.

[0180] Figure 6 The following is a schematic diagram of the physical structure of a functional surface generation device for an intersection, such as Figure 6 As shown, the intersection functional surface generation device may include: a processor (processor) 810, a communication interface (CommunicationsInterface) 820, a memory (memory) 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the intersection functional surface generation method.

[0181] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0182] On the other hand, the present application also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the intersection functional surface generation method provided by the above-mentioned methods.

[0183] On the other hand, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the intersection functional surface generation method provided by the above-mentioned methods.

[0184] It should be noted that the technical solutions or technical features described in the above embodiments can be combined or supplemented with each other without causing conflicts. The scope of protection of this application is not limited to the precise structures described in the above embodiments and shown in the drawings; all modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of this application should be included in the scope of protection of this application.

Claims

1. A method for generating a functional surface of an intersection, characterized in that: The method comprises: Based on the topological relationship of each road section of the target intersection, determining the outer edge boundary line of the adjacent road section of the target intersection; Determining a connection order of a plurality of clusters of the target intersection according to a connection relationship of each of the outer edge boundary lines, wherein the plurality of clusters are obtained by clustering features of a road section of the target intersection; Connecting the vectorized stop line of the target intersection with the outer boundary line according to the connection order of the plurality of clusters to generate an intersection function surface corresponding to the target intersection; The step of determining the outer edge boundary line of the adjacent road segments of the target intersection based on the topological relationship of the road segments of the target intersection includes: Determining a road surface type of the target intersection based on a topological relationship of the target intersection; According to the road surface type, an outer boundary line from an entrance to an exit of an adjacent road section in the target intersection is determined.

2. The method for generating a functional surface of an intersection according to claim 1, characterized in that: The outer boundary line includes a boundary line corresponding to a turning trajectory and / or a straight trajectory of the vehicle.

3. The method for generating a functional surface of an intersection according to claim 2, characterized in that: Determining the outer edge boundary line from the entrance to the exit of the adjacent road section in the target intersection according to the road surface type includes: If the road surface type is a crossroad, determining the right turning trajectory from the entrance to the intersection to the exit of each section of the target intersection as the outer edge boundary line; If the road surface type is a T-shaped intersection, determine the right turning trajectory from the entrance to the exit of each section of the target intersection and the straight track from the entrance to the exit of the section where the right turning trajectory is empty as the outer edge boundary line; If the road surface type is a straight intersection, it is determined that the straight trajectory corresponding to the empty section of the right turn trajectory of the target intersection from the entrance to the exit is the outer boundary line.

4. The method for generating a functional surface of an intersection according to claim 1, characterized in that: The step of determining the outer boundary line from the entrance to the exit of the adjacent road section in the target intersection according to the road surface type includes: Performing perception vectorization processing on the target intersection to obtain lane information of the current road section; Determine the rightmost lane in the road section entering the intersection according to the lane information; The outer edge dividing line of the right turn and / or straight driving of the rightmost lane is determined according to the road surface type as the outer edge boundary line.

5. The method for generating a functional surface of an intersection as claimed in claim 4, characterized in that: After performing the perceptual vectorization processing on the target intersection, the method further includes: The acquired vectorized road section is cut off using a stop line or a crosswalk to obtain lane information of the current road section after the cut off.

6. The method for generating a functional surface of an intersection according to claim 1, characterized in that: Before determining the connection order of the plurality of clusters at the target intersection according to the connection relationship of each of the outer boundary lines, the method further includes: Constructing a virtual area including the center point of the target intersection; Clustering is performed on the features of the intersections of the road sections in the target intersection that intersect the virtual area to generate the plurality of clusters.

7. The method for generating a functional surface of an intersection according to claim 6, characterized in that: After clustering the intersections of the virtual area and the road sections related to the target intersection to generate the plurality of clusters, the method further includes: In response to the number of the plurality of clusters being less than a threshold, the clusters associated with the target intersection are supplemented according to a symmetry principle so that the number of the plurality of clusters is greater than or equal to the threshold.

8. The method for generating a functional surface of an intersection according to claim 1, characterized in that: The step of determining the outer boundary line of the adjacent road segment of the target intersection includes: The following processing is performed on each of the multiple clusters: Determining the type of the cluster according to the identifier of the road segment corresponding to the cluster; In response to the type of the cluster being a supplementary cluster, determining a boundary line of a straight track from an entrance to an exit from an intersection corresponding to the supplementary cluster as the outer boundary line; In response to the type of the cluster being a real cluster, a boundary line of a turning trajectory from an entrance to an exit from an intersection corresponding to the real cluster is determined as the outer boundary line.

9. A device for generating a functional surface of an intersection, characterized in that: The method comprises a memory, a processor and a road intersection functional surface generation program stored in the memory and executable on the processor. When the processor executes the road intersection functional surface generation program, the steps of the road intersection functional surface generation method as described in any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a junction functional surface generation program, which, when executed, implements the steps of the junction functional surface generation method according to any one of claims 1 to 8.

Images