OSM map construction method and system for unmanned driving

CN118913239APending Publication Date: 2024-11-08BEIJING MECHANICAL EQUIP INST
-1 Cites -1 Cited by

Patent Information

Application Number
CN202310510421.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-08
Publication Date
2024-11-08

Smart Images

  • Figure CN118913239A_ABST
    Figure CN118913239A_ABST
Patent Text Reader

Abstract

The invention relates to an OSM map construction method and system for unmanned driving, belongs to the technical field of map construction, and solves the problem that a map cannot be automatically constructed in an existing scene in which use of a laser radar is prohibited. Comprising the following steps: acquiring attributes of each section of road and track points acquired on one lane in each section of road; according to the coordinates of the track points, the limit extension distance of every three adjacent track points is calculated in sequence, and according to the limit extension distance and the width of the lane where the track points are located, noisy points are removed to obtain reference track points; performing uniform resampling on a lane where the reference track point is located, and calculating coordinates of a sampling point; taking the sampling point as a point on the center line of the lane, respectively calculating expansion points on two sides of the sampling point, taking the expansion points as boundary points of the lane, and for a road with a plurality of lanes, sequentially expanding all lane boundary points of the road according to the boundary points of the lane; and converting each lane boundary point into an OSM data format, and constructing an OSM map. The map is automatically constructed based on the track points.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of map construction, and in particular to an OSM map construction method and system for unmanned driving. Background Art

[0002] OSM (OpenStreetMap) is a navigation map for unmanned driving and has been widely used. Due to its open source nature, it is welcomed by a large number of users. As an open source map format, OSM map allows users to create their own OSM maps while following the map format.

[0003] Currently, the methods for producing OSM maps include manual and semi-automatic. Manual production uses special OSM map editing software, while semi-automatic production uses sensors such as lidar and combined navigation to build high-definition maps. At the same time, with the support of technologies such as vision-based lane recognition, road-level high-definition maps are built. On this basis, the high-definition maps are converted into OSM maps and manually constructed using special software.

[0004] In many actual unmanned application scenarios, such as campus scenes and transportation scenes, which are relatively simple, they do not require complicated high-definition map composition methods. In addition, in some military scenes, laser radar scanning and scene reconstruction are not allowed, and the existing map construction method cannot adapt to the above two types of scenarios. Summary of the invention

[0005] In view of the above analysis, the embodiments of the present invention aim to provide an OSM map construction method and system for unmanned driving, so as to solve the existing problem that maps cannot be automatically constructed in scenarios where the use of lidar is prohibited.

[0006] On the one hand, an embodiment of the present invention provides an OSM map construction method for unmanned driving, comprising the following steps:

[0007] Obtain the attributes of each road segment and the trajectory points collected on a lane in each road segment;

[0008] According to the coordinates of the trajectory points, the limit extension distance of each three adjacent trajectory points is calculated in turn, and according to the limit extension distance and the width of the lane where the trajectory points are located, the noise points in the trajectory points are removed to obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points;

[0009] According to the attributes of each section of the road, the sampling point is taken as the point on the center line of the lane, and the extension points on both sides of the sampling point are calculated as the boundary points of the lane. For roads with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points.

[0010] Convert each lane boundary point into OSM data format and construct an OSM map.

[0011] Based on the further improvement of the above method, according to the coordinates of the trajectory points, the limit extension distance of each three adjacent trajectory points is calculated in sequence, including:

[0012] According to the coordinates of every three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line formed by every two adjacent trajectory points are calculated;

[0013] According to the first trajectory point of each of the three adjacent trajectory points, a third straight line perpendicular to the first straight line is constructed on one side; according to the limit extension distance, a fourth straight line parallel to the first straight line and a fifth straight line parallel to the second straight line are constructed on the same side; based on the point normal equations of the straight lines, the straight line equations of the third straight line, the fourth straight line and the fifth straight line are constructed respectively;

[0014] According to the coordinates of each three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line, when the third straight line, the fourth straight line and the fifth straight line intersect at one point, the expansion distance in the straight line equation is solved as the limit expansion distance.

[0015] Based on the further improvement of the above method, based on the point normal equation of the straight line, the three straight line equations of the third straight line, the fourth straight line and the fifth straight line are constructed as follows:

[0016] u1(xx a )+v1(yy a )=0

[0017] -v1(xx a +v1d)+u1(yy a -u1d)=0

[0018] -v2(xx c +v2d)+u2(yy c -u2d)=0

[0019] Among them, (u1, v1) represents the unit direction vector of the first straight line, (u2, v2) represents the unit direction vector of the second straight line, (x a ,y a ) represents the coordinates of the first trajectory point, (x c ,y c ) represents the coordinates of the third trajectory point, and d represents the extension distance.

[0020] Based on the further improvement of the above method, the noise points in the trajectory points are removed according to the limit extension distance and the lane width, including:

[0021] It is determined whether the limit expansion distance corresponding to the three adjacent trajectory points is greater than half of the lane width. If so, the three adjacent trajectory points are retained. Otherwise, the second trajectory point among the three adjacent trajectory points is used as a noise point and the noise points are removed from the trajectory points.

[0022] Based on the further improvement of the above method, the lane where the reference trajectory point is located is uniformly resampled according to the resampling frequency, and the coordinates of the sampling point are calculated according to the coordinates of the reference trajectory point, including:

[0023] The cumulative length of each reference track point is calculated by taking the sum of the cumulative length of the first point and the distance between the two points as the cumulative length of the second point;

[0024] According to the sampling frequency and the total length of the trajectory, the total number of sampling points and the cumulative length of the sampling points are obtained; each sampling point is taken out in turn, and according to the cumulative length of each sampling point and the reference trajectory point, the adjacent reference trajectory points before and after each sampling point are obtained;

[0025] The coordinates of each sampling point are calculated based on the cumulative length of each sampling point and the cumulative lengths and coordinates of the adjacent reference trajectory points before and after it.

[0026] Based on further improvement of the above method, the road attributes include: the total number of lanes, the direction of each lane, the width of each lane, the lane number of the collected trajectory point, the number of lanes in the same direction as the lane of the collected trajectory point, and the number of lanes in the opposite direction.

[0027] Based on the further improvement of the above method, the sampling point is taken as the point on the center line of the lane, and the extension points on both sides of the sampling point are calculated as the boundary points of the lane, including:

[0028] Along the heading direction, every two adjacent sampling points form a trajectory line, and the unit normal vector of each trajectory line is calculated; half of the lane width is used as the extension distance;

[0029] Along the two unit normal vector directions of the starting trajectory line passing through the starting sampling point, the starting extension points on both sides are obtained according to the extension distance; along the two unit normal vector directions of the last trajectory line passing through the ending sampling point, the ending extension points on both sides are obtained according to the extension distance;

[0030] For other sampling points, on both sides of the trajectory line, according to the unit normal vector of each trajectory line, an extension line is constructed that is parallel to each trajectory line and separated from the corresponding trajectory line by the extension distance. The intersection point of each two adjacent extension lines is taken as the middle extension point on both sides;

[0031] The starting extension point, the middle extension point and the ending extension point on each side constitute the boundary points on each side of the lane.

[0032] Based on the further improvement of the above method, for a road with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points, including:

[0033] Take the lane you are in as the known lane and perform the following expansion steps: according to the attributes of the road you are in, obtain the lane to be expanded that is adjacent to the known lane; according to the position of the lane to be expanded and the resampling frequency, uniformly resample the boundaries between the known lane and the lane to be expanded; calculate the coordinates of the boundary sampling points based on the coordinates of the lane boundary points; use the width of the lane to be expanded as the expansion distance, calculate the expansion points of the boundary sampling points, and use them as the boundary points of the lane to be expanded; take the lane to be expanded as the known lane, and perform the expansion steps again until there are no lanes to be expanded adjacent to the known lanes, and obtain all lane boundary points.

[0034] Based on the further improvement of the above method, each lane boundary point is converted into OSM data format to construct an OSM map, including:

[0035] Based on the node object in the OSM data format, each lane boundary point is constructed into a node object; based on the route object in the OSM data format, each lane boundary is constructed into a route object, and each route object contains the node object of each lane boundary point on the corresponding lane boundary; based on the relationship object in the OSM data format, each two adjacent lane boundaries are constructed into a relationship object, and each relationship object contains the route objects of the corresponding two adjacent lane boundaries;

[0036] Write node objects, route objects and relationship objects into OSM files, read OSM files, and build OSM maps.

[0037] On the other hand, an embodiment of the present invention provides an OSM map construction system for unmanned driving, including:

[0038] A trajectory point collection module is used to obtain the attributes of each road section and the trajectory points collected on a lane in each road section;

[0039] The resampling module is used to calculate the limit extension distance of each three adjacent trajectory points in turn according to the coordinates of the trajectory points, remove the noise points in the trajectory points according to the limit extension distance and the lane width, and obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points;

[0040] The lane boundary point expansion module is used to calculate the expansion points on both sides of the sampling point as the points on the center line of the lane according to the attributes of each section of the road. The expansion points are used as the lane boundary points. For roads with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points.

[0041] The map generation module is used to convert the lane boundary points into OSM data format and construct an OSM map.

[0042] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects: by identifying the limit extension distance, removing the noise points in the original trajectory, and then performing uniform resampling based on the denoised trajectory, the accuracy of the road boundary is ensured and the smoothness of the road boundary is improved. It is suitable for quickly and automatically constructing maps in simple scenes and military scenes where laser radar scanning is prohibited; the method for denoising the original trajectory is simple, which is not only suitable for the construction of the map of this scheme, but also suitable for unmanned driving tracking.

[0043] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0045] Figure 1 This is a flow chart of an OSM map construction method for unmanned driving in Embodiment 1 of the present invention;

[0046] Figure 2 Schematic diagram of the trajectory point denoising method in Example 1 of the present invention. DETAILED DESCRIPTION

[0047] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0048] Example 1

[0049] A specific embodiment of the present invention discloses an OSM map construction method for unmanned driving, such as Figure 1 As shown, the following steps are included:

[0050] S11. Obtaining attributes of each road section and trajectory points collected on a lane in each road section.

[0051] It should be noted that the road trajectory points are collected by the user through remote control of the drone by human-computer interaction. Considering that the road is generally divided into straight segments and curved segments, and the road includes one or more lanes, during the collection, the road information is input through the human-computer interaction interface, and the trajectory data of one lane of each road section is collected in sections according to the characteristics of different road sections (such as curvature, width, speed limit). Preferably, in order to facilitate the subsequent generation of lane boundaries, the outermost lane of each road section is selected along the direction of movement.

[0052] Specifically, the combined inertial navigation equipment of the UAV uses RTK (Real Time Kinematic) technology to feedback positioning point data at a fixed frequency (for example, 100hz). In order to avoid invalid data, the moving distance of the positioning point will be detected. Whenever the positioning distance exceeds 0.2m, a positioning point will be taken as a trajectory point. The data of the collected trajectory points include: longitude, latitude and heading angle. The heading angle is usually the angle with due north; the attributes of each road section include: the total number of lanes, the direction of each lane, the width of each lane, whether each lane is two-way or one-way, the lane number of the collected trajectory point, the number of lanes in the same direction as the lane of the collected trajectory point, and the number of lanes in the opposite direction. The collected trajectory point data and the corresponding road attribute information are stored.

[0053] S12. According to the coordinates of the trajectory points, the limit extension distance of each three adjacent trajectory points is calculated in turn, and according to the limit extension distance and the width of the lane where the trajectory points are located, the noise points in the trajectory points are removed to obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points.

[0054] It should be noted that this embodiment uses the collected trajectory points as points on the center line of the lane, and when the lane boundary is extended to both sides, the extension distance is half of the lane width. Due to the existence of noise points, a "fallback" phenomenon may occur when generating the road boundary.

[0055] like Figure 2As shown, A, B, and C are three trajectory points. Draw a straight line perpendicular to AB from point A, and draw a straight line perpendicular to BC from point C. In the first case, according to the extension distance, point A1 and point C1 are extended respectively; then draw a parallel line to AB through point A1, and draw a parallel line to BC through point C1. The intersection point B1 of the two parallel lines is the extension point of B. If A is the starting point of the entire trajectory, then A1 is the extension point of A. At this time, the extended point B1 is after point A1, and there is no "backward" phenomenon.

[0056] When the extension distance is relatively large, or the angle between AB and BC is relatively small, for example, in the second case, points A2 and C2 are extended. At this time, parallel lines are drawn based on the extended A2 and C2. The intersection point B2 of the two parallel lines coincides with A2, that is, the two parallel lines of A2 and C2 intersect with the straight line from A to A2 at one point. This situation does not conform to the actual situation.

[0057] When the extension distance continues to increase, in the third case, the intersection point B3 of the two parallel lines, as the extension point of point B, is in front of the extension point A3 of point A, and a "fallback" phenomenon occurs, which is also not in line with the actual situation.

[0058] Therefore, the above-mentioned case ② is the critical case of expansion, and the distance from A to A2 is the limit expansion distance. When the cases ② and ③ occur, that is, the expansion distance is greater than or equal to the limit expansion distance, the trajectory point B is a noise point.

[0059] Specifically, according to the coordinates of the trajectory points, the limit extension distance of each three adjacent trajectory points is calculated, including:

[0060] According to the coordinates of every three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line formed by every two adjacent trajectory points are calculated;

[0061] According to the first trajectory point of each of the three adjacent trajectory points, a third straight line perpendicular to the first straight line is constructed on one side; according to the limit extension distance, a fourth straight line parallel to the first straight line and a fifth straight line parallel to the second straight line are constructed on the same side; based on the point normal equation of the straight line, the three straight line equations of the third straight line, the fourth straight line and the fifth straight line are constructed respectively as shown below:

[0062] u1(xx a )+v1(yy a )=0 formula (1)

[0063] -v1(xx a +v1d)+u1(yy a -u1d)=0 Formula (2)

[0064] -v2(xx c +v2d)+u2(yyc -u2d)=0 Formula (3)

[0065] Among them, (u1, v1) represents the unit direction vector of the first straight line, (u2, v2) represents the unit direction vector of the second straight line, (x a ,y a ) represents the coordinates of the first trajectory point, (x c ,y c ) represents the coordinates of the third trajectory point, and d represents the extension distance;

[0066] According to the coordinates of each three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line, when the third straight line, the fourth straight line and the fifth straight line intersect at one point, the extension distance in the straight line equation is solved as the limit extension distance, and (x, y) is obtained as the intersection coordinates.

[0067] It should be noted that, according to the unit direction vectors of the first straight line and the second straight line, it can be known that the unit normal vector of the third straight line is (u1, v1), the unit normal vector of the fourth straight line is (-v1, u1), and the unit normal vector of the fifth straight line is (-v2, u2).

[0068] It is worth noting that, based on the above analysis and the calculation method of the limit extension, the process of removing noise points in this step is an iterative process. Except for the starting point and the end point, one trajectory point is taken out in turn as the current trajectory point, and the current trajectory point and its front and rear adjacent trajectory points are taken as the current three adjacent trajectory points. The limit extension distance is calculated using the straight line equation, and then based on the comparison between the limit extension distance and the extension distance, it is judged whether the limit extension distance corresponding to the current three adjacent trajectory points is greater than half of the lane width. If greater, the current three adjacent trajectory points are retained, otherwise, the current trajectory point, that is, the second trajectory point among the three adjacent trajectory points, is taken as a noise point, and the noise point is removed from the trajectory points; continue to take out the next trajectory point and its front and rear adjacent trajectory points as the current three adjacent trajectory points, and perform noise point judgment until the judgment of all trajectory points is completed, and the remaining trajectory points are the reference trajectory points.

[0069] Exemplarily, the collected trajectory points are A, B, C, D and E. The starting point A and the end point E do not need to be judged. B is taken out first, and B and its front and back adjacent A and C are used as the current three adjacent trajectory points to determine whether B is a noise point. If B is a noise point, it is removed. Then C is taken out. Since B in front of C has been removed, its front and back adjacent are A and D. Then, based on A, C, and D as the current three adjacent trajectory points, it is determined whether point C is a noise point. If C is not a noise point, then C, D, and E are used as the current three adjacent trajectory points to determine whether point D is a noise point. If D is a noise point, it is removed. Finally, A, C, and E are obtained as the reference trajectory points.

[0070] Compared with the prior art, the method for denoising the original trajectory in this embodiment is simple, and is not only applicable to the construction of the map of this solution, but also to the tracking of unmanned driving.

[0071] Furthermore, since the denoised reference trajectory points are uneven, in order to improve the smoothness of the road boundary, the lanes where the reference trajectory points are located are uniformly resampled according to the resampling frequency, and the coordinates of the sampling points are calculated based on the coordinates of the reference trajectory points.

[0072] Specifically, the following steps are included:

[0073] The cumulative length of each reference track point is calculated by taking the sum of the cumulative length of the first point and the distance between the two points as the cumulative length of the second point;

[0074] According to the sampling frequency, the cumulative length of each sampling point is obtained; each sampling point is taken out in turn, and the adjacent reference trajectory points before and after each sampling point are obtained according to the cumulative length of each sampling point and the reference trajectory point;

[0075] According to the cumulative length of each sampling point, and the cumulative length and coordinates of the adjacent reference trajectory points before and after it, the coordinates of each sampling point are calculated by the following formula:

[0076]

[0077] Among them, (x i ,y i ) represents the coordinates of the i-th sampling point, i = 1, 2, ..., L / f, L represents the total length of the trajectory, f represents the sampling frequency, l i =i×f represents the cumulative length of the i-th sampling point, (x i1 ,y i1 )、(x i2 ,y i2 ) represent the coordinates of the two adjacent reference trajectory points before and after the i-th sampling point, l i1 , l i2 They respectively represent the cumulative lengths of the two adjacent reference trajectory points before and after the i-th sampling point.

[0078] Compared with the prior art, this embodiment performs uniform resampling based on the denoised trajectory, thereby ensuring the accuracy of the road boundary and improving the smoothness of the road boundary.

[0079] S13. According to the attributes of each section of the road, the sampling point is taken as the point on the center line of the lane, and the extension points on both sides of the sampling point are calculated respectively as the boundary points of the lane. For a road with multiple lanes, all lane boundary points of the road are expanded in sequence according to the boundary points of the lane.

[0080] It should be noted that this step is a multi-layer iterative process. First, each section of the road is taken out in turn for processing. For each road, half of the lane width of the sampling point is used as the expansion distance to iteratively generate the lane boundary point where the sampling point is located. Then, uniform resampling is performed based on the lane boundary points. The obtained boundary sampling points use the adjacent lane width of the lane as the expansion distance, and the adjacent lane boundary points are expanded outward based on the boundary sampling points. Then, the expansion continues outward based on the adjacent lane boundary points until all lane boundary points are obtained.

[0081] Specifically, for each road section, the sampling point is taken as the point on the center line of the lane, and the extension points on both sides of the sampling point are calculated as the boundary points of the lane, including:

[0082] Along the heading direction, every two adjacent sampling points form a trajectory line, and the unit normal vector of each trajectory line is calculated; half of the lane width is used as the extension distance;

[0083] Along the two unit normal vector directions of the starting trajectory line passing through the starting sampling point, the starting extension points on both sides are obtained according to the extension distance; along the two unit normal vector directions of the last trajectory line passing through the ending sampling point, the ending extension points on both sides are obtained according to the extension distance;

[0084] For other sampling points, on both sides of the trajectory line, according to the unit normal vector of each trajectory line, an extension line is constructed that is parallel to each trajectory line and separated from the corresponding trajectory line by the extension distance. The intersection point of each two adjacent extension lines is taken as the middle extension point on both sides;

[0085] The starting extension point, the middle extension point and the ending extension point on each side constitute the boundary points on each side of the lane.

[0086] Preferably, the intersection of the extension lines is calculated by constructing a straight line equation; if the initial trajectory point is the outermost lane selected along the direction of movement, the outermost boundary of the outermost lane is generated first, and then the other side boundary of the outermost lane is generated.

[0087] Specifically, for a road with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points, including:

[0088] Take the lane you are in as the known lane and perform the following expansion steps: according to the attributes of the road you are in, obtain the lane to be expanded that is adjacent to the known lane; according to the position of the lane to be expanded and the resampling frequency, uniformly resample the boundaries between the known lane and the lane to be expanded; calculate the coordinates of the boundary sampling points based on the coordinates of the lane boundary points; use the width of the lane to be expanded as the expansion distance, calculate the expansion points of the boundary sampling points, and use them as the boundary points of the lane to be expanded; take the lane to be expanded as the known lane, and perform the expansion steps again until there are no lanes to be expanded adjacent to the known lanes, and obtain all lane boundary points.

[0089] For example, the initial trajectory point is the rightmost lane selected along the direction of movement, and the adjacent lane to be expanded is on its left. After obtaining the rightmost lane boundary, the left boundary of the rightmost lane is uniformly resampled to obtain the left boundary sampling point, and the adjacent lane to be expanded is expanded to the left according to the left boundary sampling point; then determine whether there is another lane on the left side of the adjacent lane to be expanded, and if so, continue to expand. If not, complete all lane boundary points of the road.

[0090] It should be noted that when generating lane boundary points, the direction of the lane is also recorded to facilitate more accurate map generation.

[0091] S14. Convert each lane boundary point into an OSM data format to construct an OSM map.

[0092] It should be noted that the OSM data format mainly includes three basic objects: nodes, ways and relations, which constitute the entire map. Among them, nodes define the location of points in space; ways define routes or areas; relations define the relationship between objects. In addition, attribute information is added to various objects through tags.

[0093] In this embodiment, based on the node object in the OSM data format, each lane boundary point is constructed into a node object; based on the route object in the OSM data format, each lane boundary is constructed into a route object, and each route object contains the node object of each lane boundary point on the corresponding lane boundary, which is used to represent the relationship between the lane boundary and the node; based on the relationship object in the OSM data format, each two adjacent lane boundaries are constructed into a relationship object, and each relationship object contains the route objects of the corresponding two adjacent lane boundaries, which are used to represent the relationship between the lanes;

[0094] Write node objects, route objects and relationship objects into OSM files, read OSM files, and build OSM maps.

[0095] Compared with the prior art, the present embodiment provides an OSM map construction method for unmanned driving, which removes noise points in the original trajectory by identifying the extreme extension distance, and then performs uniform resampling based on the denoised trajectory, thereby ensuring the accuracy of the road boundary and improving the smoothness of the road boundary. The method is suitable for quickly and automatically constructing maps in simple scenes and military scenes where laser radar scanning is prohibited. The method for denoising the original trajectory is simple, which is not only suitable for the construction of the map of this solution, but also suitable for unmanned driving tracking.

[0096] Example 2

[0097] Another embodiment of the present invention discloses an OSM map construction system for unmanned driving, thereby realizing an OSM map construction method for unmanned driving in embodiment 1. The specific implementation of each module refers to the corresponding description in embodiment 1, and the system includes:

[0098] A trajectory point collection module is used to obtain the attributes of each road section and the trajectory points collected on a lane in each road section;

[0099] The resampling module is used to calculate the limit extension distance of each three adjacent trajectory points in turn according to the coordinates of the trajectory points, remove the noise points in the trajectory points according to the limit extension distance and the lane width, and obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points;

[0100] The lane boundary point expansion module is used to calculate the expansion points on both sides of the sampling point as the points on the center line of the lane according to the attributes of each section of the road. The expansion points are used as the lane boundary points. For roads with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points.

[0101] The map generation module is used to convert the lane boundary points into OSM data format and construct an OSM map.

[0102] Since the OSM map construction system for unmanned driving in this embodiment and the OSM map construction method for unmanned driving described above can be mutually referenced, it is a repeated description here, so it will not be repeated here. Since the principles of this system embodiment and the above method embodiment are the same, this system embodiment also has the corresponding technical effects of the above method embodiment.

[0103] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0104] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for constructing an OSM map for unmanned driving, characterized in that: The following steps are involved: Obtain the attributes of each road segment and the trajectory points collected on a lane in each road segment; According to the coordinates of the trajectory points, the limit extension distance of each three adjacent trajectory points is calculated in turn, and according to the limit extension distance and the width of the lane where the trajectory points are located, the noise points in the trajectory points are removed to obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points; According to the attributes of each section of the road, the sampling point is taken as the point on the center line of the lane, and the extension points on both sides of the sampling point are calculated as the boundary points of the lane. For roads with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points. Convert each lane boundary point into OSM data format and construct an OSM map.

2. The method for constructing an OSM map for unmanned driving according to claim 1, characterized in that: The step of sequentially calculating the limit extension distances of three adjacent trajectory points according to the coordinates of the trajectory points includes: According to the coordinates of every three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line formed by every two adjacent trajectory points are calculated; According to the first trajectory point of each of the three adjacent trajectory points, a third straight line perpendicular to the first straight line is constructed on one side; according to the limit extension distance, a fourth straight line parallel to the first straight line and a fifth straight line parallel to the second straight line are constructed on the same side; based on the point normal equations of the straight lines, the straight line equations of the third straight line, the fourth straight line and the fifth straight line are constructed respectively; According to the coordinates of each three adjacent trajectory points, the unit direction vectors of the first straight line and the second straight line, when the third straight line, the fourth straight line and the fifth straight line intersect at one point, the expansion distance in the straight line equation is solved as the limit expansion distance.

3. The method for constructing an OSM map for unmanned driving according to claim 2, characterized in that: The three straight line equations of the third straight line, the fourth straight line and the fifth straight line are constructed based on the point normal equation of the straight line as shown below: u1(x-x a )+v1(y-y a )=0 -v1(xx a +v1d)+u1(yy a -u1d)=0 <h2 style=";text-align:left;direction:ltr">-v2(xx<h2 style=";text-align:left;direction:ltr"> c <h2 style=";text-align:left;direction:ltr"> +v2d)+u2(yy<h2 style=";text-align:left;direction:ltr"> c <h2 style=";text-align:left;direction:ltr"> -u2d)=0 Among them, (u1, v1) represents the unit direction vector of the first straight line, (u2, v2) represents the unit direction vector of the second straight line, (x a ,y a ) represents the coordinates of the first trajectory point, (x c ,y c ) represents the coordinates of the third trajectory point, and d represents the extension distance.

4. The method for constructing an OSM map for unmanned driving according to claim 2, characterized in that: The removing of noise points in the track points according to the limit extension distance and the lane width includes: It is determined whether the limit expansion distance corresponding to the three adjacent trajectory points is greater than half of the lane width. If so, the three adjacent trajectory points are retained. Otherwise, the second trajectory point among the three adjacent trajectory points is used as a noise point, and the noise point is removed from the trajectory points.

5. The method for constructing an OSM map for unmanned driving according to claim 1, characterized in that: The method of uniformly resampling the lane where the reference trajectory point is located according to the resampling frequency and calculating the coordinates of the sampling point according to the coordinates of the reference trajectory point includes: The cumulative length of each reference track point is calculated by taking the sum of the cumulative length of the first point and the distance between the two points as the cumulative length of the second point; According to the sampling frequency and the total length of the trajectory, the total number of sampling points and the cumulative length of the sampling points are obtained; each sampling point is taken out in turn, and according to the cumulative length of each sampling point and the reference trajectory point, the adjacent reference trajectory points before and after each sampling point are obtained; The coordinates of each sampling point are calculated based on the cumulative length of each sampling point and the cumulative lengths and coordinates of the adjacent reference trajectory points before and after it.

6. The method for constructing an OSM map for unmanned driving according to claim 1, characterized in that: The road attributes include: the total number of lanes, the direction of each lane, the width of each lane, the lane number of the collected trajectory point, the number of lanes in the same direction as the lane of the collected trajectory point, and the number of lanes in the opposite direction.

7. The method for constructing an OSM map for unmanned driving according to claim 6, characterized in that: The method of taking the sampling point as a point on the center line of the lane and calculating the extension points on both sides of the sampling point as the boundary points of the lane includes: Along the heading direction, every two adjacent sampling points form a trajectory line, and the unit normal vector of each trajectory line is calculated; half of the lane width is used as the extension distance; Along the two unit normal vector directions of the starting trajectory line passing through the starting sampling point, the starting extension points on both sides are obtained according to the extension distance; along the two unit normal vector directions of the last trajectory line passing through the ending sampling point, the ending extension points on both sides are obtained according to the extension distance; For other sampling points, on both sides of the trajectory line, according to the unit normal vector of each trajectory line, an extension line is constructed that is parallel to each trajectory line and separated from the corresponding trajectory line by the extension distance. The intersection point of each two adjacent extension lines is taken as the middle extension point on both sides; The starting extension point, the middle extension point and the ending extension point on each side constitute the boundary points on each side of the lane.

8. The method for constructing an OSM map for unmanned driving according to claim 7, characterized in that: For a road with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points, including: Take the lane you are in as the known lane and perform the following expansion steps: according to the attributes of the road you are in, obtain the lane to be expanded that is adjacent to the known lane; according to the position of the lane to be expanded and the resampling frequency, uniformly resample the boundaries between the known lane and the lane to be expanded; calculate the coordinates of the boundary sampling points based on the coordinates of the lane boundary points; use the width of the lane to be expanded as the expansion distance, calculate the expansion points of the boundary sampling points, and use them as the boundary points of the lane to be expanded; take the lane to be expanded as the known lane, and perform the expansion steps again until there are no lanes to be expanded adjacent to the known lanes, and obtain all lane boundary points.

9. The method for constructing an OSM map for unmanned driving according to claim 1, characterized in that: The step of converting each lane boundary point into the OSM data format and constructing an OSM map includes: Based on the node object in the OSM data format, each lane boundary point is constructed into a node object; based on the route object in the OSM data format, each lane boundary is constructed into a route object, and each route object contains the node object of each lane boundary point on the corresponding lane boundary; based on the relationship object in the OSM data format, each two adjacent lane boundaries are constructed into a relationship object, and each relationship object contains the route objects of the corresponding two adjacent lane boundaries; Write node objects, route objects and relationship objects into OSM files, read OSM files, and build OSM maps.

10. An OSM map construction system for unmanned driving, characterized in that: include: A trajectory point collection module is used to obtain the attributes of each road section and the trajectory points collected on a lane in each road section; The resampling module is used to calculate the limit extension distance of each three adjacent trajectory points in turn according to the coordinates of the trajectory points, remove the noise points in the trajectory points according to the limit extension distance and the lane width, and obtain the reference trajectory points; according to the resampling frequency, the lane where the reference trajectory points are located is uniformly resampled, and the coordinates of the sampling points are calculated according to the coordinates of the reference trajectory points; The lane boundary point expansion module is used to calculate the expansion points on both sides of the sampling point as the points on the center line of the lane according to the attributes of each section of the road. The expansion points are used as the lane boundary points. For roads with multiple lanes, all lane boundary points of the road are expanded in sequence according to the lane boundary points. The map generation module is used to convert the lane boundary points into OSM data format and construct an OSM map.