Map construction method and system, method for formulating navigation strategy and storage medium
Through the image acquisition and positioning system, a high-precision map is built and navigation strategies are formulated, which solves the problems of high-cost and complex algorithms in traditional technologies and realizes efficient navigation capabilities of autonomous vehicles.
Patent Information
- Application Number
- CN202010265150.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-04-07
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2040-04-07
AI Technical Summary
Traditional high-precision map production requires high-cost sensor configurations and complex algorithms, making it difficult to effectively build high-precision lane lines and point cloud processing, limiting the automatic up-and-off ramp and parallel functions of autonomous driving vehicles on complex roads.
The lane line is obtained by using an image acquisition device, the positioning system is used to obtain the latitude and longitude coordinates and heading, and the polynomial fits the lane center line under the vehicle coordinate system, perform coordinate conversion, smoothing processing and equidistant points, build a high-precision map, and use it according to Map navigation strategies.
It has realized efficient construction of high-precision maps and navigation strategies, reduced sensor costs and map processing time, improved the navigation capabilities of autonomous vehicles on complex roads, and the method is simple and the calculation volume is low.
Smart Images

Figure CN113494917B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of map navigation, and in particular to a map construction method, a map construction method system, a method for formulating a navigation strategy, and a computer-readable storage medium. Background Art
[0002] High-precision maps play a core role in the field of unmanned driving. They can help cars perceive complex information on the road in advance, so that they can make correct decisions in combination with intelligent path planning. For example, on the highway, if an autonomous vehicle wants to realize functions such as automatic on- and off-ramps and automatic lane merging, it needs a high-precision map module to provide lane-level navigation information, including the current lane of the vehicle, the target lane of the vehicle, the speed limit of each lane, the drivable distance of each lane, and the variable lane interval of each lane, etc. The production of traditional high-precision maps requires sensors such as cameras, lidars, and high-precision combined inertial navigation, and the cost of sensor configuration is relatively high; in addition, the algorithms involved in traditional lane line extraction and point cloud processing are also difficult. Summary of the invention
[0003] The present invention proposes a mechanism for efficiently constructing maps and formulating navigation strategies based on them. Specifically:
[0004] According to one aspect of the present invention, a map construction method is provided, comprising the following steps: using an image acquisition device to acquire lane lines on both sides of a lane of a road and using a positioning system to acquire corresponding latitude and longitude coordinates and headings; using an M-order polynomial in a vehicle coordinate system to fit the lane lines on both sides respectively, and thereby fitting the various order coefficients of the lane center line, wherein M≥3; dividing the road into a plurality of sections, and in each section: converting the latitude and longitude coordinates into coordinates in a rectangular coordinate system; determining the coordinates of the center line in the rectangular coordinate system according to the coordinates in the rectangular coordinate system, the position of the image acquisition device in the vehicle coordinate system, the various order coefficients and the heading, wherein the vehicle coordinate system takes the positioning system as the origin; smoothing the coordinates of the center line in the rectangular coordinate system; taking points at equal distances on the smoothed center line to form data for the lane; and translating the lane to simulate data for other lanes.
[0005] Optionally, in some embodiments of the present invention, the zero-order coefficient of the center line is the average of the zero-order coefficients of the fitting polynomials of the lane lines on both sides.
[0006] Optionally, in some embodiments of the present invention, the coordinates of the center line in the rectangular coordinate system are (x_lane, y_lane), and x_lane = x_offset*cos(phi) - (A0+y_offset)*sin(phi) + x_gps, y_lane = x_offset*sin(phi) + (A0+y_offset)*cos(phi) + y_gps; wherein (x_gps, y_gps) are the coordinates in the rectangular coordinate system, A0 is the zero-order coefficient of the center line, (x_offset, y_offset) are the coordinates of the image acquisition device in the vehicle coordinate system, and phi is the heading.
[0007] Optionally, in some embodiments of the present invention, the step of smoothing the coordinates of the center line in the rectangular coordinate system includes: selecting k waypoints on both sides of the N waypoints to be smoothed, and using these N + 2*k waypoints to participate in fitting to achieve smoothing of the N waypoints, and if there are insufficient waypoints, using acceptable waypoints to participate in fitting.
[0008] Optionally, in some embodiments of the present invention, the step of taking points at equal distances on the smoothed center line to form data for the lane includes: taking a point at every distance D, and if there is no waypoint at the point where the point is taken, using interpolation to generate the point at that point.
[0009] Optionally, in some embodiments of the present invention, the step of translating the lane to simulate data of other lanes includes: determining the width of the lane based on coefficients of various orders of the fitting polynomials of the lane lines on both sides, and translating the lane according to the width of the lane to simulate data of other lanes.
[0010] According to another aspect of the present invention, there is provided a method for formulating a navigation strategy based on a map constructed using any one of the map construction methods described above, comprising the following steps: setting the attributes of each lane of each section of a road; determining in which section of the road a vehicle is currently located; determining the specific position of the vehicle in the section; determining the direction of travel of the vehicle; determining the lane in which the vehicle is located; extracting a local map based on the section where the vehicle is currently located; and formulating a navigation strategy based on the local map.
[0011] Optionally, in some embodiments of the present invention, the attributes include the number of speed limit intervals, the position of the speed limit intervals in the road section and their speed limit range, the number of variable lane intervals, the range of the variable lane intervals and the connection relationship between the road sections.
[0012] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and wherein when the instructions are executed by a processor, the processor is caused to execute any one of the methods described above.
[0013] According to another aspect of the present invention, a map construction system is provided, the system comprising: an image acquisition unit, configured to acquire lane lines on both sides of a lane of a road; an inertial navigation positioning unit, configured to acquire corresponding longitude and latitude coordinates and heading; a processing unit, configured to: fit the lane lines on both sides respectively using an M-order polynomial in a vehicle coordinate system, and thereby fit coefficients of each order of the lane center line, wherein M≥3; divide the road into a plurality of sections, and in each section: convert the longitude and latitude coordinates into coordinates in a rectangular coordinate system; determine the coordinates of the center line in the rectangular coordinate system according to the coordinates in the rectangular coordinate system, the position of the image acquisition device in the vehicle coordinate system, the coefficients of each order and the heading, wherein the vehicle coordinate system takes the positioning system as the origin; smooth the coordinates of the center line in the rectangular coordinate system; take points at equal distances on the smoothed center line to form data of the lane; and translate the lane to simulate data of other lanes. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The above and other objects and advantages of the present invention will become more fully apparent from the following detailed description taken in conjunction with the accompanying drawings, wherein the same or similar elements are denoted by the same reference numerals.
[0015] Figure 1 A schematic diagram of a map construction method according to an embodiment of the present invention is illustrated.
[0016] Figure 2 A schematic diagram of a map construction method according to an embodiment of the present invention is illustrated.
[0017] Figure 3 A schematic diagram of a map construction method according to an embodiment of the present invention is illustrated.
[0018] Figure 4 A schematic diagram of a method for formulating a navigation strategy according to an embodiment of the present invention is illustrated.
[0019] Figure 5 A schematic diagram of a method for formulating a navigation strategy according to an embodiment of the present invention is illustrated.
[0020] Figure 6 A schematic diagram of a map construction system according to an embodiment of the present invention is illustrated.
[0021] Figure 7A schematic diagram of a method for formulating a navigation strategy according to an embodiment of the present invention is illustrated. DETAILED DESCRIPTION
[0022] According to one aspect of the present invention, a map construction method is provided. Figure 1 As shown, the method includes the following steps. In step S101, the lane lines on both sides of a lane of the road are obtained by using an image acquisition device and the corresponding longitude and latitude coordinates and heading are obtained by using a positioning system; the lane lines on both sides are fitted respectively by using an M-order polynomial in the vehicle coordinate system, and the coefficients of each order of the lane center line are fitted thereby, where M≥3. In some examples of the present invention, waypoint collection is performed in step S101, which can be specifically implemented in the following manner: manually driving a vehicle and driving along the center of the lane to collect data. For a road containing two or more lanes, two lanes can be collected, one of which is used for mapping and the other is used for map verification. The collected data includes the lane lines on both sides identified by the camera, and are fitted in the vehicle coordinate system according to, for example, a cubic polynomial, and the coefficients after fitting are collected; on the other hand, the longitude, latitude and heading of the vehicle output by the high-precision GPS need to be collected. For example, the coefficients of the cubic polynomial are A3, A2, A1, and A0 from high order to low order, and then the coefficient A0 of the lane center line can be calculated by the coefficients A0_L and A0_R of the lane lines on the left and right sides. The calculation formula is A0 = (A0_L+A0_R) / 2. Figure 1 In step S101, the above process is abbreviated as "waypoint, coordinate and heading acquisition". In some examples, the image acquisition device and high-precision GPS in the present invention can be a camera and a high-precision inertial navigation system on a mass-produced vehicle, which are lower in cost than the traditional use of laser radar, camera and high-precision inertial navigation system. At the same time, the method used is simple and can effectively save time for map processing.
[0023] The method divides the road into multiple sections in step S102. Figure 1 In the example of the present invention, the road is divided into several sections according to the number of lanes, and the collected waypoints are extracted by section. In this way, each section can be processed separately, thereby reducing the data size of a single processing.
[0024] The method performs coordinate conversion in step S103, converting the longitude and latitude coordinates into coordinates in a rectangular coordinate system. Some examples of the present invention can select a suitable map origin in step S103, and convert the longitude and latitude (lon, lat) data of the original road point into coordinate points (x, y) in a plane rectangular coordinate system. The rectangular coordinate system here can be formulated as needed, for example, taking the starting position of the vehicle as the origin, the due east direction as the positive direction of the x-axis, and the due north direction as the positive direction of the y-axis. It should be noted that the rectangular coordinate system selected here will also be used in subsequent steps (sometimes also referred to as an absolute coordinate system, and the coordinates in this coordinate system are also referred to as absolute coordinates). If necessary, a rectangular coordinate system can also be formulated for each road section to avoid the expansion of cumulative errors.
[0025] The method performs waypoint correction in step S104, that is, the coordinates of the center line in the rectangular coordinate system are determined according to the coordinates in the rectangular coordinate system, the position of the image acquisition device in the vehicle coordinate system, the coefficients of each order, and the heading, wherein the vehicle coordinate system takes the vehicle-mounted positioning system as the origin. The purpose of this step is to eliminate the lateral deviation of the waypoints collected during the manual driving of the vehicle relative to the lane center.
[0026] In some embodiments of the present invention, the zero-order coefficient of the center line is the average of the zero-order coefficients of the fitting polynomials of the lane lines on both sides. In some embodiments of the present invention, the coordinates of the center line in the rectangular coordinate system are (x_lane, y_lane), and
[0027] x_lane = x_offset*cos(phi) - (A0+y_offset)*sin(phi) + x_gps,
[0028] y_lane = x_offset*sin(phi) + (A0+y_offset)*cos(phi) + y_gps;
[0029] (x_gps, y_gps) are the coordinates in the rectangular coordinate system, the zero-order coefficient of the A0 centerline, (x_offset, y_offset) are the coordinates of the image acquisition device in the vehicle coordinate system, and phi is the heading.
[0030] Specifically, in some examples of the present invention, in step S104, the lane centerline coefficient A0 obtained in step S101 can be used to correct the waypoint. The correction idea is: according to the known GPS point and the relative relationship between the GPS and the lane centerline, the GPS point is corrected to the lane centerline. Figure 2As shown, A is the GPS position, B is the camera position, and C is the lane centerline position identified by the camera. In the vehicle coordinate system with point A as the origin, the coordinates of point B in the vehicle coordinate system are (x_offset, y_offset), and the position of point C relative to point B in the vehicle coordinate system is (0, A0). The vehicle heading angle phi, based on the absolute coordinates of GPS point A (x_gps, y_gps), the absolute coordinates of the lane centerline C (x_lane, y_lane) are obtained.
[0031] First, we can get the coordinate of point C in the vehicle coordinate system as (x_offset, A0+y_offset). At this time, according to the absolute coordinate of point A, the absolute coordinate of point C can be obtained through coordinate system conversion. The conversion formula is as follows:
[0032] x_lane = x_offset*cos(phi) - ( A0+y_offset)*sin(phi) + x_gps;
[0033] y_lane = x_offset*sin(phi) + (A0+y_offset)*cos(phi) + y_gps.
[0034] In step S105, the method performs smoothing processing on the coordinates of the center line in the rectangular coordinate system ( Figure 1 In some examples of the present invention, considering the discontinuity problem of A0 between two adjacent frames output by the camera, which will cause the waypoints to jump after the deviation correction, it is necessary to smooth the waypoints after the deviation correction. The smoothing method may be, for example, fitting the points after the deviation correction in step S104. For example, in some embodiments of the present invention, the step of smoothing the coordinates of the center line in the rectangular coordinate system includes: selecting k waypoints on both sides of the N waypoints to be smoothed, and involving these N + 2*k waypoints in fitting to achieve smoothing of the N waypoints, and if there are insufficient waypoints, involving the available waypoints in fitting. Specifically, for roads with relatively long distances, in order to avoid the problem of large errors caused by one-time fitting of the entire section, a segmented fitting method is adopted. The idea of segmented fitting is as follows. Figure 3 As shown in the figure, N points are selected as the segments to be smoothed each time. However, in order to ensure that two adjacent segments are relatively continuous after smoothing, k additional points are taken at the beginning and end of the segment to be smoothed, for a total of N+2*k points for fitting. If there are less than k points before and after the segment to be smoothed, such as the beginning and end of the data, the actual number of points can be extracted.
[0035] In step S106, the method takes points at equal distances on the smoothed center line to form the data of the lane ( Figure 1In the figure, it is abbreviated as "equidistant waypoint extraction"). After smoothing, GPS waypoints are relatively dense, occupy a large storage space, and are not equally spaced. In order to reduce the map storage space and facilitate subsequent use, the GPS points are sparsely processed. The extraction method is such as taking a point every distance D by interpolation method, and finally completing the map data processing of a single lane. For example, in some embodiments of the present invention, the step of taking points at equal distances on the center line after smoothing to form the data of the lane includes: taking a point every distance D, and if there is no waypoint at the point, using interpolation method to generate the point at that point.
[0036] The method in step S107 translates the lane to simulate data of other lanes ( Figure 1 In some embodiments of the present invention, the step of translating the lane to simulate the data of other lanes includes: determining the width of the lane according to the coefficients of each order of the fitting polynomial of the lane lines on both sides, and translating the lane according to the width of the lane to simulate the data of other lanes. Specifically, for a road section with multiple lanes, first calculate the width of the lane according to the coefficients A0 and A1 of the lane lines on both sides extracted in step S101, and then translate the map data of the single lane processed in step S106 to obtain other lanes according to the lane width. The processed data of each road section are stored separately. In some examples of the present invention, the lane IDs are accumulated from 1 from left to right.
[0037] According to another aspect of the present invention, a method for formulating a navigation strategy based on a map constructed using any one of the above map construction methods is provided, thereby constructing a lane-level navigation function. Figure 7 As shown, the method includes the following steps.
[0038] The method sets the attributes of each lane of each road section in step S701 ( Figure 7In the present invention, the attributes refer to various intelligence information about the road that can be obtained, which is used to provide the driver with information during manual driving, including but not limited to information that can be obtained in a visual form. In some embodiments of the present invention, the attributes include but are not limited to the number of speed limit intervals of each lane of each road section, the position of the speed limit interval in the road section and its speed limit range, the number of variable lane intervals, the range of variable lane intervals, and the connection relationship between each road section. Specifically, in step S701, the road attributes can be defined, for example, the number of speed limit intervals of each lane of each road section can be customized; the speed limit interval of each section of each lane of each road section can be customized, including the distance from the end point of the speed limit interval to the starting point of the road section, the minimum speed limit and the maximum speed limit; the number of variable lane intervals of each lane of each road section can be customized; the range of each variable lane interval of each lane of each road section can be customized, including the distance from the starting point of the road section to the end point of the variable lane interval; the road connection relationship between the current road section and the previous road section can be defined according to the actual road conditions; the target lane ID of each road section can be customized to generate the navigation path of the vehicle.
[0039] The method determines in step S702 which section of the road the vehicle is currently located on ( Figure 7 In some embodiments of the present invention, the current road segment ID is determined in step S702. A navigation path may contain many road segments. It is necessary to first determine the road segment ID of the vehicle according to the vehicle coordinates, and then retrieve the map of the corresponding road segment. The method for determining the road segment ID is as follows: Figure 4 As shown. Taking road section 2 as an example, first use an envelope rectangle to enclose road section 2, and at the same time draw a straight line 1 along the road normal direction at the junction of road section 2 and road section 1, and draw a straight line 2 along the road normal direction at the junction of road section 2 and road section 3. If the vehicle coordinates are in the rectangular frame and between straight line 1 and straight line 2 at the same time, it indicates that the vehicle is on road section 2.
[0040] The method determines the specific position of the vehicle in the road section in step S703 ( Figure 7 In some embodiments of the present invention, the vehicle's neighboring points are retrieved to determine the position in step S703. In step S702, the road section ID is determined, and the map data of the road section is retrieved. Next, the position of the vehicle coordinates in the map data is calculated, that is, it is necessary to retrieve the two points Pf and Pr in front and behind the vehicle in the map data. During initialization, that is, during the first retrieval, for straight roads, a binary search is used, and for curves, a traversal method is used. After the retrieval is completed, the index numbers Pf_index and Pr_index of the two points in the map data are saved. After the initialization is completed, the subsequent retrieval uses the last retrieval result as the initial value, and searches before and after it until the two neighboring points in front and behind the vehicle are found.
[0041] The method determines the driving direction of the vehicle in step S704 ( Figure 7 In some examples of the present invention, according to the two front and rear neighboring points Pf and Pr retrieved in step S703, the angle between the vector PrPf and the vehicle heading in the absolute coordinate system is calculated. If the absolute value of the angle is less than 90 degrees, the vehicle's driving direction is defined as forward, otherwise, the vehicle's driving direction is reverse (relative to the "front and rear" defined above).
[0042] The method determines the lane in which the vehicle is located in step S705 ( Figure 7 In some examples of the present invention, the current lane ID is determined in step S705: if the road section where the vehicle is located has many lanes, the lane ID of the vehicle is determined according to the vehicle position. The front and rear adjacent points of the vehicle have been calculated before. A local coordinate system is established with these two points. Pr is the origin of the coordinate system, the vector PrPf is the x-axis direction, and the vertical direction on the left of the vector PrPf is the positive y-axis direction. The lateral coordinate of the vehicle position in the local coordinate system is calculated, and then the lane ID of the vehicle is determined according to the lane width.
[0043] In step S706, the method extracts a local map according to the road section where the vehicle is currently located ( Figure 7 Different navigation paths contain different numbers of road segments, and different road segments contain different numbers of lanes. In order to make the lane-level navigation module more versatile, the local map around the vehicle is abstracted, such as Figure 5 As shown, the local map only includes 4 road segments: the vehicle's current road segment, the vehicle's rear road segment, the vehicle's front road segment 1, and the vehicle's front road segment 2. According to the road segment ID calculated in step S702, the map data of the 4 road segments, the number of lanes, and the road attributes of the 4 road segments defined in step S701 are extracted each time.
[0044] Determine the connection relationship between each lane of other sections and each lane of the current section based on the extracted road attributes. Define the maximum number of lanes as 10, and define the connection relationship as a 1*10 array Junc. For example, Junc[2] = 3 means that the second lane of this section is connected to the third lane of the current section. Figure 5 As shown, the previous road section has only one lane, which is connected to the third lane of the current road section. Therefore, the connection relationship between the previous road section and the current road section is [3,0,0,0,0,0,0,0,0,0]; similarly, the connection relationships between the vehicle's front section 1 and the vehicle's front section 2 and the current road section are [0,1,2,0,0,0,0,0,0,0,0] and [2,0,0,0,0,0,0,0,0,0,0] respectively.
[0045] The method formulates a navigation strategy based on the local map in step S707 ( Figure 7 In some examples of the present invention, navigation information processing is performed in step S707. After the local map of the vehicle's surroundings is extracted, based on the local map and the road connection relationship between road sections, lane-level navigation information is extracted and given to the planning control module. The extraction of lane-level navigation information mainly includes: extracting the variable lane section in front of the vehicle; extracting the speed limit section in front of the vehicle; extracting the vehicle's drivable distance; extracting the front and rear waypoints of the vehicle and the number of valid front and rear waypoints. The extracted front and rear waypoints are then transferred to the vehicle coordinate system, and a navigation strategy can be formulated based on this.
[0046] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein instructions are stored in the computer-readable storage medium, and the characteristics are that when the instructions are executed by a processor, the processor is caused to execute any one of the methods described above. The computer-readable medium referred to in the present invention includes various types of computer storage media, which can be any available medium that can be accessed by a general-purpose or special-purpose computer. For example, the computer-readable medium can include RAM, ROM, EPROM, E 2 PROM, register, hard disk, removable disk, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or can be used to carry or store the desired program code unit in the form of instruction or data structure and can be accessed by general or special purpose computer, or general or special purpose processor. Any other temporary or non-temporary medium. As used herein, disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and blue-ray disc, wherein the disk usually copies data magnetically, and the disc uses laser to copy data optically. The above combination should also be included in the protection scope of computer-readable medium. Exemplary storage medium is coupled to the processor so that the processor can read and write information from / to the storage medium. In an alternative, the storage medium can be integrated into the processor. The processor and storage medium can reside in the ASIC. The ASIC can reside in the user terminal. In an alternative, the processor and storage medium can reside in the user terminal as discrete components.
[0047] According to another aspect of the present invention, a map construction system is provided. Figure 7 As shown, the system 60 includes an image acquisition unit 602 , an inertial navigation positioning unit 604 , and a processing unit 606 .
[0048] In some examples of the present invention, the image acquisition unit 602 is configured to obtain lane lines on both sides of a lane of the road. The inertial navigation positioning unit 604 is configured to obtain the corresponding longitude and latitude coordinates and heading. The processing unit 606 is configured to: fit the lane lines on both sides respectively using an M-order polynomial in the vehicle coordinate system, and thereby fit the various order coefficients of the lane center line, where M≥3; divide the road into multiple sections, and in each section: convert the longitude and latitude coordinates into coordinates in a rectangular coordinate system; determine the coordinates of the center line in the rectangular coordinate system according to the coordinates in the rectangular coordinate system, the position of the image acquisition device in the vehicle coordinate system, the various order coefficients and the heading, where the vehicle coordinate system takes the positioning system as the origin; smooth the coordinates of the center line in the rectangular coordinate system; take points at equal distances on the smoothed center line to form the data of the lane; and translate the lane to simulate the data of other lanes.
[0049] In some examples of the present invention, the specific details of the various tasks performed by the image acquisition unit 602, the inertial navigation positioning unit 604 and the processing unit 606 may be further restricted as described in other embodiments above, and will not be repeated here.
[0050] In view of the above, some examples of the present invention propose a method for constructing a high-precision map and navigation function for the development and testing of autonomous driving. Based on the road to be tested, a high-precision map and lane-level navigation function can be quickly constructed to support the development and testing of autonomous driving functions. Compared with the products provided by traditional map vendors, the cost is lower and the efficiency and flexibility are higher. In addition, the sensors used in some examples of the present invention: mass-produced car cameras and high-precision inertial navigation are lower in cost than the traditional use of lidar, cameras and high-precision inertial navigation. At the same time, the method used is simple and can effectively save the time of map processing. Finally, some examples of the present invention propose a method for constructing a lane-level navigation function, abstracting the local map around the vehicle, and then extracting navigation information based on the local map. The method is simple to construct and has low computational complexity. At the same time, the proposed method of abstracting the local map of the vehicle makes the lane-level navigation module more versatile.
[0051] The above examples mainly illustrate the map construction method, map construction method system, method for formulating navigation strategy and computer-readable storage medium of the present invention. Although only some of the embodiments of the present invention are described, it should be understood by those skilled in the art that the present invention can be implemented in many other forms without departing from its subject matter and scope. Therefore, the examples and embodiments shown are regarded as illustrative rather than restrictive, and the present invention may cover various modifications and substitutions without departing from the spirit and scope of the present invention as defined in the appended claims.
Claims
1. A map construction method, characterized in that: The method comprises the following steps: Using an image acquisition device to obtain lane lines on both sides of a lane of the road and using a positioning system to obtain corresponding latitude and longitude coordinates and heading; In the vehicle coordinate system, the lane lines on both sides are fitted respectively using an M-order polynomial, and the coefficients of each order of the lane center line are fitted thereby, wherein M≥3; Divide the road into multiple segments, and under each segment: Convert the latitude and longitude coordinates into coordinates in a rectangular coordinate system; Determine the coordinates of the center line in the rectangular coordinate system according to the coordinates in the rectangular coordinate system, the position of the image acquisition device in the vehicle coordinate system, the coefficients of each order, and the heading, wherein the vehicle coordinate system takes the positioning system as the origin; Smoothing the coordinates of the center line in the rectangular coordinate system; The smoothed center line is equidistantly selected to form the data of the lane; as well as Move the lane to simulate the data of other lanes. Wherein, the coordinates of the center line in the rectangular coordinate system are (x_lane, y_lane), and x_lane=x_offset*cos(phi)-(A0+y_offset)*sin(phi)+x_gps, y_lane=x_offset*sin(phi)+(A0+y_offset)*cos(phi)+y_gps; Wherein (x_gps, y_gps) are the coordinates in the rectangular coordinate system, A0 is the zero-order coefficient of the center line, (x_offset, y_offset) are the coordinates of the image acquisition device in the vehicle coordinate system, and phi is the heading.
2. The method according to claim 1, characterized in that The zero-order coefficient of the center line is the average of the zero-order coefficients of the fitting polynomials of the lane lines on both sides.
3. The method according to claim 1, characterized in that The step of smoothing the coordinates of the center line in the rectangular coordinate system includes: selecting k waypoints on both sides of the N waypoints to be smoothed, and using these N+2*k waypoints to participate in fitting to achieve smoothing of the N waypoints, and if there are insufficient waypoints, using available waypoints to participate in fitting.
4. The method according to claim 1, characterized in that: The step of taking points at equal distances on the smoothed center line to form data of the lane includes: taking a point at every distance D, and if there is no waypoint at the taking point, using interpolation to generate the point at that location.
5. The method according to claim 1, characterized in that The step of translating the lane to simulate data of other lanes includes: determining the width of the lane according to coefficients of each order of the fitting polynomial of the lane lines on both sides, and translating the lane according to the width of the lane to simulate data of other lanes.
6. A method for formulating a navigation strategy based on a map constructed using the map construction method according to any one of claims 1 to 5, characterized in that: The method for formulating a navigation strategy comprises the following steps: Set the properties of each lane of each road section; Determining which section of the road the vehicle is currently located on; Determine the specific location of the vehicle on the road section; Determine the direction of the vehicle; Determine the lane the vehicle is in; Extracting a local map based on the road section the vehicle is currently on; and A navigation strategy is formulated based on the local map.
7. The method according to claim 6, characterized in that The attributes include the number of speed limit sections, the position of the speed limit sections in the road section and the speed limit range, the number of variable lane sections, the range of the variable lane sections and the connection relationship between the road sections.
8. A computer-readable storage medium, wherein instructions are stored in the computer-readable storage medium, characterized in that: When the instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 7.
9. A computer program product, characterized in that The computer program product comprises instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 7.
10. A map construction system, characterized in that: The system comprises: An image acquisition unit configured to acquire lane lines on both sides of a lane of a road; An inertial navigation positioning unit configured to obtain corresponding latitude and longitude coordinates and heading; A processing unit configured to: In the vehicle coordinate system, the lane lines on both sides are fitted respectively using an M-order polynomial, and the coefficients of each order of the lane center line are fitted thereby, wherein M≥3; Divide the road into multiple segments, and under each segment: Convert the latitude and longitude coordinates into coordinates in a rectangular coordinate system; Determine the coordinates of the center line in the rectangular coordinate system according to the coordinates in the rectangular coordinate system, the position of the image acquisition unit in the vehicle coordinate system, the coefficients of each order, and the heading, wherein the vehicle coordinate system takes the inertial navigation positioning unit as the origin; Smoothing the coordinates of the center line in the rectangular coordinate system; The smoothed center line is equidistantly selected to form the data of the lane; and Move the lane to simulate the data of other lanes. Wherein, the coordinates of the center line in the rectangular coordinate system are (x_lane, y_lane), and x_lane=x_offset*cos(phi)-(A0+y_offset)*sin(phi)+x_gps, y_lane=x_offset*sin(phi)+(A0+y_offset)*cos(phi)+y_gps; Wherein (x_gps, y_gps) are the coordinates in the rectangular coordinate system, A0 is the zero-order coefficient of the center line, (x_offset, y_offset) are the coordinates of the image acquisition unit in the vehicle coordinate system, and phi is the heading.
