A method and apparatus for constructing a road map

Abstract

The application discloses a kind of construction method and device of road map, obtain real lane line data and real inertial navigation data under vehicle coordinate system, to determine the lane line discrete point information belonging to each lane line respectively under global coordinate system to obtain target lane line data, target lane line data is divided into reference line lane line data and non-reference line lane line data, and is respectively handled to obtain target reference line lane line data and target lane data, based on target reference line lane line data, the inertial navigation data of car under global coordinate system is handled to obtain target road height data, target reference line lane line data, target lane data and target road height data are constructed according to OpenDRIVE protocol to obtain OpenDRIVE road map.The OpenDRIVE road map established by real lane line data and real inertial navigation data in the application has high reliability, and the accuracy of intelligent driving simulation test result can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent driving, more particularly, to a road map construction method and device. BACKGROUND

[0002] An intelligent driving system is an important application of artificial intelligence in the transportation field, and intelligent driving simulation testing is an important step in the design and implementation process of an intelligent driving system. In the design process of intelligent driving, a road map, as a carrier for vehicle travel, describes a static traffic scene in which a vehicle under test is located, and is an important component of intelligent driving simulation testing.

[0003] In the prior art, when intelligent driving simulation testing is performed, a road map is usually manually built by a builder in combination with a testing purpose, a functional specification and the builder's own experience. Therefore, the road map is a manual implementation of the builder's idea, and there is a large gap between the road map and a real road map. The intelligent driving simulation testing result cannot accurately test the situation of an intelligent driving algorithm under a real road map, thereby resulting in a certain error in the intelligent driving simulation testing result. SUMMARY

[0004] Therefore, the present application discloses a road map construction method and device, so that the OpenDRIVE road map constructed has high reliability, can accurately test the situation of an intelligent driving algorithm under a real road map, is suitable for the field of intelligent driving simulation testing, and can greatly improve the accuracy of the intelligent driving simulation testing result.

[0005] A road map construction method comprises the following steps:

[0006] Real lane line data and real inertial navigation data in a vehicle coordinate system are obtained.

[0007] Lane line discrete point information respectively belonging to each lane line in a global coordinate system is determined based on the real lane line data and the real inertial navigation data, and target lane line data is obtained.

[0008] A global lane line is selected as a reference lane line from the target lane line data, and the target lane line data is divided into reference line lane line data and non-reference line lane line data based on the reference lane line.

[0009] The reference line lane line data is processed by using a preset reference line processing scheme, and target reference line lane line data is obtained.

[0010] The non-reference line lane line data is processed by using a preset non-reference line processing scheme, and target lane data is obtained.

[0011] processing the inertial navigation data of the vehicle in the global coordinate system based on the target reference line lane line data, wherein the inertial navigation data of the vehicle is obtained by converting the real inertial navigation data from the latitude and longitude coordinate system to the global coordinate system;

[0012] generating road map related parameters according to the OpenDRIVE protocol based on the target reference line lane line data, the target lane data and the target road height data, and constructing an OpenDRIVE road map.

[0013] Optionally, the target lane line data is determined based on the real lane line data and the real inertial navigation data in the global coordinate system, and includes:

[0014] a preset number of discrete points are uniformly extracted from the lane line curve corresponding to the real lane line data, to obtain first lane line data;

[0015] the lane line discrete point coordinates in the vehicle coordinate system are converted into lane line discrete point coordinates in the global coordinate system based on the first lane line data and the global pose data of the vehicle coordinate origin, to obtain second lane line data;

[0016] the second lane line data is clustered by using a clustering algorithm to obtain the target lane line data.

[0017] Optionally, the reference line lane line data is processed by using a preset reference line processing scheme to obtain target reference line lane line data, including:

[0018] the reference lane line corresponding to the reference line lane line data is uniformly and interval divided into a plurality of segments, and each segment contains a plurality of lane line discrete points in the global coordinate system, to obtain first intermediate reference lane line data;

[0019] a straight line is fitted by using a least square method for the lane line discrete points in each segment of the first intermediate reference lane line data, and a midpoint of the straight line is selected, to obtain second intermediate reference lane line data;

[0020] a B-spline curve is obtained by fitting a cubic for the lane line discrete points in the second intermediate reference lane line data, and curve parameters of the B-spline curve are determined, to obtain third intermediate reference lane line data;

[0021] the parameter equation of each curve in the third intermediate reference lane line data in the global coordinate system is converted into the local coordinate system, to obtain the target reference line lane line data.

[0022] Optionally, the non-reference line lane data is processed by using a preset non-reference line processing scheme to obtain target lane data, including:

[0023] All lane line discrete points in the non-reference line lane data are converted from a global coordinate system to a reference coordinate to obtain first intermediate non-reference line lane data;

[0024] The first intermediate non-reference line lane data is uniformly segmented along the s-axis direction of the reference line coordinate system, and each segment contains a plurality of lane line discrete points in the reference coordinate system to obtain second intermediate non-reference line lane data;

[0025] The lane line discrete points in each segment in the second intermediate non-reference line lane data are linearly fitted by using a least square method, and points are taken along the s-axis of the reference line coordinate system to obtain third intermediate non-reference line lane data;

[0026] Based on the definition of lane numbering in OpenDRIVE, the lane number corresponding to each lane line in the third intermediate non-reference line lane data is determined to obtain fourth intermediate non-reference line lane data;

[0027] Based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane data, the lane width corresponding to each lane line discrete point is determined to obtain fifth intermediate non-reference line lane data;

[0028] When the number of lanes changes each time, the fifth intermediate non-reference line lane data is re-divided into lane sections based on the definition of lane section in OpenDRIVE to obtain sixth intermediate non-reference line lane data;

[0029] The lane width in the sixth intermediate non-reference line lane data is fitted by using a least square method to obtain the target lane data.

[0030] Optionally, the vehicle inertial navigation data in the global coordinate system is processed based on the target reference line lane data to obtain target road height data, including:

[0031] Based on the target reference line lane data, the vehicle inertial navigation position information corresponding to all timestamps in the vehicle inertial navigation data in the global coordinate system is converted to the reference line coordinate system to obtain initial road height data;

[0032] The initial road height data is fitted by using a least square method to obtain target road height data.

[0033] Optionally, the target reference line lane line data, the target lane data and the target road height data are generated into road map related parameters according to the OpenDRIVE protocol, and an OpenDRIVE road map is constructed, comprising:

[0034] The target reference line lane line data is generated into reference line description in the OpenDRIVE road map according to the OpenDRIVE protocol;

[0035] The target lane data RLD8 is generated into road description in the OpenDRIVE road map according to the OpenDRIVE protocol;

[0036] The target road height data is generated into road height description in the OpenDRIVE road map according to the OpenDRIVE protocol.

[0037] A road map construction device, comprising:

[0038] An acquisition unit configured to acquire real lane line data and real inertial navigation data in a vehicle coordinate system;

[0039] A lane line data determination unit configured to determine lane line discrete point information respectively belonging to each lane line in a global coordinate system based on the real lane line data and the real inertial navigation data, and obtain target lane line data;

[0040] A lane line data division unit configured to select a global lane line as a reference lane line from the target lane line data, and divide the target lane line data into reference line lane line data and non-reference line lane line data based on the reference lane line;

[0041] A first data processing unit configured to process the reference line lane line data using a preset reference line processing scheme, and obtain target reference line lane line data;

[0042] A second data processing unit configured to process the non-reference line lane line data using a preset non-reference line processing scheme, and obtain target lane data;

[0043] A third data processing unit configured to process vehicle inertial navigation data in the global coordinate system based on the target reference line lane line data, and obtain target road height data, wherein the vehicle inertial navigation data is obtained by converting the real inertial navigation data from latitude and longitude coordinates to the global coordinate system;

[0044] A map construction unit configured to generate road map related parameters from the target reference line lane line data, the target lane data and the target road height data according to the OpenDRIVE protocol, and construct an OpenDRIVE road map.

[0045] Optionally, the lane line data determination unit specifically comprises:

[0046] a discrete point extraction subunit, configured to extract a preset number of discrete points from the lane line curve corresponding to the real lane line data at uniform intervals, to obtain first lane line data;

[0047] a first coordinate system conversion subunit, configured to convert the lane line discrete point coordinates in the vehicle coordinate system into lane line discrete point coordinates in the global coordinate system based on the first lane line data and global pose data of the vehicle coordinate origin, to obtain second lane line data;

[0048] a clustering subunit, configured to cluster the second lane line data by using a clustering algorithm to obtain the target lane line data.

[0049] Optionally, the first data processing unit specifically comprises:

[0050] a first segmentation subunit, configured to uniformly and interval segment the reference lane line corresponding to the reference line lane line data into a plurality of segments, each of the segments containing a plurality of lane line discrete points in the global coordinate system, to obtain first intermediate reference lane line data;

[0051] a first fitting subunit, configured to fit a straight line to the lane line discrete points in each segment of the first intermediate reference lane line data by using a least square method, and select a midpoint of the straight line, to obtain second intermediate reference lane line data;

[0052] a second fitting subunit, configured to fit a B-spline curve by dividing the lane line discrete points in the second intermediate reference lane line data into segments, and determine curve parameters of the B-spline curve, to obtain third intermediate reference lane line data;

[0053] a second coordinate system conversion subunit, configured to convert the parametric equation of each curve in the third intermediate reference lane line data in the global coordinate system to the local coordinate system, to obtain the target reference line lane line data.

[0054] Optionally, the second data processing unit specifically comprises:

[0055] a third coordinate system conversion subunit, configured to convert all lane line discrete points in the non-reference line lane line data from the global coordinate system to the reference coordinate, to obtain first intermediate non-reference line lane line data;

[0056] The second segmentation subunit is used to divide the lane lines corresponding to the first intermediate non-reference line lane line data into multiple segments at uniform intervals along the s-axis direction of the reference line coordinate system. Each segment contains multiple lane line discrete points under the reference coordinate system to obtain the second intermediate non-reference line lane line data.

[0057] The third fitting subunit is used to perform least squares linear fitting on the discrete points of the lane lines in the reference coordinate system in each segment of the second intermediate non-reference line lane line data, and to take points at uniform intervals along the s-axis of the reference line coordinate system to obtain the third intermediate non-reference line lane line data.

[0058] The lane number determination subunit is used to determine the lane number corresponding to each lane line in the third intermediate non-reference line lane line data based on the definition of lane number in OpenDRIVE, so as to obtain the fourth intermediate non-reference line lane line data.

[0059] The lane width determination subunit is used to determine the lane width corresponding to each lane line discrete point based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane line data, and to obtain the fifth intermediate non-reference line lane line data.

[0060] The lane line data partitioning subunit is used to re-divide the lane line data of the fifth intermediate non-reference line based on the definition of lane section in OpenDRIVE each time the number of lanes changes, to obtain the lane line data of the sixth intermediate non-reference line.

[0061] The fourth fitting subunit is used to fit the lane width in the sixth intermediate non-reference line lane line data piecewise using the least squares method to obtain the target lane data.

[0062] As can be seen from the above technical solution, the present invention discloses a method and apparatus for constructing a road map. The method involves acquiring real lane line data and real inertial navigation data in a vehicle coordinate system; determining the discrete point information of each lane line belonging to a specific lane line in the global coordinate system based on the real lane line data and the real inertial navigation data to obtain target lane line data; selecting a global lane line from the target lane line data as a reference lane line; dividing the target lane line data into reference lane line data and non-reference lane line data based on the reference lane line; processing the reference lane line data using a preset reference line processing scheme to obtain target reference lane line data; processing the non-reference lane line data using a preset non-reference line processing scheme to obtain target lane data; processing the vehicle's inertial navigation data in the global coordinate system based on the target reference lane line data to obtain target road height data; and generating road map-related parameters using the target reference lane line data, target lane data, and target road height data according to the OpenDRIVE protocol to construct an OpenDRIVE road map. As can be seen, this invention uses real lane line data and real inertial navigation data from visual perception to build an OpenDRIVE road map. Since the visual perception results of each frame are based on the vehicle's own coordinate system, the constructed OpenDRIVE road map has high reliability and can accurately test the performance of intelligent driving algorithms under real road maps. It is applicable to fields such as intelligent driving simulation testing and can greatly improve the accuracy of intelligent driving simulation test results. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the published drawings without creative effort.

[0064] Figure 1 This is a flowchart of a road map construction method disclosed in an embodiment of the present invention;

[0065] Figure 2 This is a flowchart of a method for obtaining target lane line data in a global coordinate system, as disclosed in an embodiment of the present invention.

[0066] Figure 3 This is a flowchart of a method for obtaining target reference line lane line data according to an embodiment of the present invention;

[0067] Figure 4 This is a flowchart of a method for acquiring target lane data disclosed in an embodiment of the present invention;

[0068] Figure 5This is a flowchart of a method for obtaining road height data disclosed in an embodiment of the present invention;

[0069] Figure 6 This is a flowchart of a method for generating road map-related parameters by using target reference line lane line data, target lane data, and target road height data according to the OpenDRIVE protocol, as disclosed in an embodiment of the present invention, to construct an OpenDRIVE road map;

[0070] Figure 7 This is a schematic diagram of a road map construction device disclosed in an embodiment of the present invention. Detailed Implementation

[0071] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0072] This invention discloses a method and apparatus for constructing a road map. The method involves acquiring real lane line data and real inertial navigation data in a vehicle coordinate system. Based on the real lane line data and real inertial navigation data, discrete point information of each lane line belonging to a specific lane line in the global coordinate system is determined to obtain target lane line data. A global lane line is selected from the target lane line data as a reference lane line. Based on the reference lane line, the target lane line data is divided into reference lane line data and non-reference lane line data. The reference lane line data is processed using a preset reference line processing scheme to obtain target reference lane line data. The non-reference lane line data is processed using a preset non-reference line processing scheme to obtain target lane data. Based on the target reference lane line data, the vehicle's inertial navigation data in the global coordinate system is processed to obtain target road height data. The target reference lane line data, target lane data, and target road height data are then used to generate road map-related parameters according to the OpenDRIVE protocol to construct an OpenDRIVE road map. As can be seen, this invention uses real lane line data and real inertial navigation data from visual perception to build an OpenDRIVE road map. Since the visual perception results of each frame are based on the vehicle's own coordinate system, the constructed OpenDRIVE road map has high reliability and can accurately test the performance of intelligent driving algorithms under real road maps. It is applicable to fields such as intelligent driving simulation testing and can greatly improve the accuracy of intelligent driving simulation test results.

[0073] To facilitate understanding of the road map construction process disclosed in this invention, the following explanation is provided:

[0074] This invention uses OpenDRIVE as the road network description format. OpenDRIVE is a map format standard proposed by MSC and currently maintained by ASAM (Automotive Standardization Organization). Its data is represented using Extensible Markup Language (XML) syntax, and the file extension is ".xodr". OpenDRIVE road maps can typically be used to build road scenarios using simulation software.

[0075] This invention does not consider mapping scenarios involving road intersections, roundabouts, etc. when constructing road maps.

[0076] The coordinate system used in constructing road maps according to this invention is described below:

[0077] The vehicle coordinate system corresponds to the vehicle coordinate system in the ISO 8855 standard. The x-axis is horizontal, pointing directly in front of the vehicle and parallel to the vehicle's longitudinal plane of symmetry. The y-axis is perpendicular to the vehicle's longitudinal plane of symmetry and points to the left side of the vehicle. The z-axis is vertically upward, and the origin is fixed at a point on the vehicle.

[0078] The global coordinate system corresponds to the Inertial coordinate system in OpenDRIVE, with the x-axis pointing to the right, the y-axis pointing upwards, and the z-axis perpendicular to the drawing plane and pointing outwards; the corresponding geographic reference has the x-axis pointing east, the y-axis pointing north, and the z-axis pointing vertically upwards, with the origin fixed at a certain geographic location.

[0079] The reference line coordinate system corresponds to the reference line coordinate system in OpenDRIVE. The s-axis is along the tangential direction of the reference line, the t-axis is perpendicular to the s-axis and points to the left, and the h-axis is perpendicular to the st plane and points upward. The reference line coordinate system is positioned in the global coordinate system by providing the global coordinates of the origin and the direction angle of the s-axis of the origin in the global coordinate system.

[0080] The local coordinate system corresponds to the local coordinate system in OpenDRIVE. The u-axis corresponds to the s-axis of the reference line coordinate system, the v-axis corresponds to the t-axis, and the z-axis corresponds to the h-axis. The local coordinate system is positioned within the reference line coordinate system by providing the reference line coordinates of the origin and the orientation angle of the u-axis of the origin in the reference line coordinate system.

[0081] The specific construction process of the vehicle coordinate system, global coordinate system, reference line coordinate system, and local coordinate system can be found in existing mature solutions, and will not be elaborated here.

[0082] See Figure 1 The present invention discloses a flowchart of a method for constructing a road map, the method comprising:

[0083] Step S101: Obtain real lane line data and real inertial navigation data in the vehicle coordinate system;

[0084] In this invention, real lane line data and real inertial navigation data are acquired by a road survey vehicle, which refers to a vehicle that collects data on real roads. In this embodiment, the road survey vehicle is equipped with a visual sensor (e.g., a camera) and an inertial navigation sensor. The visual sensor collects real lane line data, and the inertial navigation sensor collects real inertial navigation data.

[0085] Real-world lane line data includes time-stamped lane line IDs and cubic polynomial coefficients of the lane lines in the vehicle coordinate system after inverse perspective transformation. In practical applications, in addition to collecting real-world lane line data, the vision sensor can also collect time-stamped traffic sign and traffic light information. Traffic light information includes: traffic light ID, traffic light type, and traffic light location.

[0086] The actual inertial navigation data, also known as the vehicle's inertial navigation pose data, includes: latitude and longitude with timestamps, altitude, vehicle yaw angle, roll angle, and pitch angle information.

[0087] Step S102: Based on real lane line data and real inertial navigation data, determine the discrete point information of each lane line belonging to each lane line in the global coordinate system to obtain the target lane line data;

[0088] Step S103: Select a global lane line from the target lane line data as a reference lane line, and divide the target lane line data into reference lane line data and non-reference lane line data based on the reference lane line.

[0089] Among them, the reference lane line data CLD1 includes the global lane line ID, which is the discrete point of the lane line in the global coordinate system.

[0090] Non-reference lane line data RLD1 includes:

[0091] Global lane line IDx: Discrete points of lane lines in the global coordinate system (including the location information of the points and lane line type, color, and width);

[0092] Global lane line IDy: Discrete points of lane lines in the global coordinate system (including the location information of the points and the lane line type, color, and width).

[0093] Step S104: Process the reference line lane line data using a preset reference line processing scheme to obtain the target reference line lane line data;

[0094] Step S105: Process the non-reference line lane data using a preset non-reference line processing scheme to obtain the target lane data;

[0095] Step S106: Based on the target reference line lane line data, process the vehicle's inertial navigation data in the global coordinate system to obtain the target road height data;

[0096] Specifically, based on the target reference line lane line data CLD5, the vehicle's inertial navigation position information corresponding to all timestamps in the global coordinate system is transformed to the reference line coordinate system, and the least squares method is used for piecewise fitting to obtain the target road height data ED2.

[0097] Among them, the inertial navigation data VD2 of this vehicle is obtained by converting the actual inertial navigation data VD1 from the latitude and longitude coordinate system to the global coordinate system.

[0098] Step S107: Generate road map related parameters by using the target reference line lane line data, target lane data and target road height data according to the OpenDRIVE protocol, and construct the OpenDRIVE road map.

[0099] The road map parameters include: reference line description, lane description, and road height description.

[0100] In summary, this invention discloses a method for constructing a road map. The method involves acquiring real lane line data and real inertial navigation data in a vehicle coordinate system. Based on the real lane line data and real inertial navigation data, it determines the discrete point information of each lane line belonging to a specific lane line in the global coordinate system to obtain target lane line data. A global lane line is selected from the target lane line data as a reference lane line. Based on the reference lane line, the target lane line data is divided into reference lane line data and non-reference lane line data. The reference lane line data is processed using a preset reference line processing scheme to obtain target reference lane line data. The non-reference lane line data is processed using a preset non-reference line processing scheme to obtain target lane data. Based on the target reference lane line data, the vehicle's inertial navigation data in the global coordinate system is processed to obtain target road height data. The target reference lane line data, target lane data, and target road height data are then used to generate road map-related parameters according to the OpenDRIVE protocol to construct an OpenDRIVE road map. As can be seen, this invention uses real lane line data and real inertial navigation data from visual perception to build an OpenDRIVE road map. Since the visual perception results of each frame are based on the vehicle's own coordinate system, the constructed OpenDRIVE road map has high reliability and can accurately test the performance of intelligent driving algorithms under real road maps. It is applicable to fields such as intelligent driving simulation testing and can greatly improve the accuracy of intelligent driving simulation test results.

[0101] To further optimize the above embodiments, see [link to relevant documentation]. Figure 2 The flowchart of a method for obtaining target lane line data in a global coordinate system disclosed in this embodiment of the invention, specifically step S102 includes:

[0102] Step S201: Extract a preset number of discrete points at uniform intervals from the lane curves corresponding to the real lane line data to obtain the first lane line data;

[0103] The real lane line data LD1 uses the lane line cubic polynomial coefficients to describe the lane line geometry information in the vehicle coordinate system after perspective transformation, that is, the lane line curve. A preset number of discrete points are extracted from the lane line curve at uniform intervals, starting from the nearest point, to obtain the first lane line data LD2.

[0104] Step S202: Based on the first lane line data and the global pose data of the vehicle coordinate origin, convert the discrete point coordinates of the lane line in the vehicle coordinate system to the discrete point coordinates of the lane line in the global coordinate system to obtain the second lane line data.

[0105] The global pose data of the vehicle's coordinate origin is obtained based on the real inertial navigation data VD1.

[0106] Specifically, the real inertial navigation data VD1 is transformed from the latitude and longitude coordinate system to the global coordinate system to obtain the vehicle's inertial navigation data VD2 in the global coordinate system; linear interpolation is performed on the vehicle's inertial navigation data VD2 in the global coordinate system to obtain the vehicle's inertial navigation data corresponding to the lane line data timestamp; based on the vehicle's inertial navigation data and the inertial navigation sensor installation parameters (installation position and installation angle), the global pose data VD3 of the vehicle's coordinate system origin at each timestamp is obtained.

[0107] The first lane line data LD2 and the global pose data VD3 are used to convert the discrete point coordinates of the lane line in the vehicle coordinate system to the discrete point coordinates of the lane line in the global coordinate system, thus obtaining the second lane line data LD3.

[0108] Step S203: Use a clustering algorithm to cluster the second lane line data to obtain the target lane line data.

[0109] Specifically, the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) clustering algorithm can be used to cluster all lane line discrete points in the global coordinate system (i.e., the second lane line data LD3). The DBSCAN clustering algorithm is a density-based clustering algorithm that can group lane line discrete point samples belonging to the same lane line into one class, obtaining the lane line discrete point information belonging to each lane line in the global coordinate system, i.e., the target lane line data LD4.

[0110] To further optimize the above embodiments, see [link to relevant documentation]. Figure 3 The flowchart of a method for obtaining target reference line lane line data disclosed in this embodiment of the invention, specifically step S104, may include:

[0111] Step S301: Divide the reference lane line corresponding to the reference lane line data into multiple segments at uniform intervals. Each segment contains multiple lane line discrete points in the global coordinate system to obtain the first intermediate reference lane line data.

[0112] In this embodiment, the reference lane line data CLD1 is segmented to obtain the first intermediate reference lane line data CLD2.

[0113] Step S302: Fit a straight line to the discrete points of the lane lines in each segment of the first intermediate reference lane line data using the least squares method, and select the midpoint of the straight line to obtain the second intermediate reference lane line data.

[0114] In this embodiment, the second intermediate reference lane line data is represented as CLD3.

[0115] Step S303: Piecewise fit the discrete points of the lane lines in the second intermediate reference lane line data three times to obtain the B-spline curve, and determine the curve parameters of the B-spline curve to obtain the third intermediate reference lane line data.

[0116] The curve parameters include: the starting coordinates of each B-spline curve in the global coordinate system, the curve length, the starting coordinates of the curve in the reference line coordinate system, and the angle of the tangent vector at the starting point of the curve.

[0117] Specifically, B-spline curves are obtained by fitting discrete points of the lane lines in the second intermediate reference lane line data CLD3 three times. A segment can be fitted with every four discrete points; for example, discrete points P1, P2, P3, and P4 can fit one segment of the curve, and discrete points P2, P3, P4, and P5 can fit the next segment. n discrete points can fit n-3 segments of the curve.

[0118] Each curve obtained by the cubic B-spline curve fitting algorithm is smooth and continuous, and the parameter coefficients of each curve in the global coordinate system can be obtained. For the parameter equation of each curve in the global coordinate system, when the parameter variable takes the value of 0, the coordinates of the starting point of each curve in the global coordinate system can be obtained.

[0119] The length of each B-spline curve segment is determined using the Gauss-Legend quadrature method.

[0120] The s-coordinate of the starting point of curve segment 1 in the reference line coordinate system is 0. The s-coordinate of the starting point of curve segment x is the sum of the lengths of the previous x-1 curve segments. Then, the s-coordinate of the starting point of each B-spline curve segment in the reference line coordinate system can be obtained.

[0121] Based on the geometric relationship between the cubic B-spline curve and the discrete points used for fitting, the direction of the tangent vector at the starting point of the curve is from the first point to the third point. For example, the direction of the tangent vector at the starting point of curve 1 is from P1 to P3. The angle between this tangent vector and the vector (1, 0) in the xy plane under the global coordinate system is the angle of the tangent vector at the starting point of the curve.

[0122] Based on the curve parameters of the B-spline curve, the third intermediate reference lane line data CLD4 is obtained.

[0123] Step S304: Transform the parametric equations of each curve segment in the third intermediate reference lane line data from the global coordinate system to the local coordinate system to obtain the target reference lane line data.

[0124] Specifically, the parametric equations of each curve segment in the third intermediate reference lane line data CLD4 are transformed from the global coordinate system to the local coordinate system. The origin of the local coordinate system for each curve segment is the starting point of that curve segment, and the u-axis direction is the tangent direction at the starting point of that curve segment. Using the global coordinates of the starting point and the tangent direction angle of the starting point for each curve segment, the transformation matrix between the local and global coordinate systems can be obtained. This transformation matrix is ​​then used to transform the parametric equations of each curve segment from the global coordinate system to the local coordinate system, resulting in the target reference lane line data CLD5.

[0125] To further optimize the above embodiments, see [link to relevant documentation]. Figure 4 The flowchart of a method for acquiring target lane data disclosed in this embodiment of the invention, specifically step S105, may include:

[0126] Step S401: Transform all lane line discrete points in the non-reference lane line data from the global coordinate system to the reference coordinate system to obtain the first intermediate non-reference lane line data;

[0127] Specifically, for any discrete point Po on the lane line, the distance expression from the discrete point to the curve segment is derived based on the parametric equation expression of each curve segment in the global coordinate system of the target reference line lane line data CLD5. The minimum value of the distance expression and the corresponding parameter variable at which the minimum value is taken are then calculated. From this, the curve segment closest to the discrete point, the closest distance, and the corresponding parameter variable q can be obtained. Using q and the parametric equation expression in the global coordinate system of the curve segment, the global coordinates of the nearest point Pc on the curve segment can be calculated. The sign of the t-coordinate of point Po in the reference line coordinate system can be determined by the cross product of the vector pointing from Pc to Po and the tangent vector of the curve at point Pc. Combined with the closest distance value, the reference line coordinate t value of point Po is obtained. Using q and the parametric equation expression in the global coordinate system of the curve segment, the curve length from the start point of the curve segment to point Pc can be calculated using the Gauss-Legend quadrature method. Adding the s-coordinate of the start point of the curve segment, the reference line coordinate s value of points Pc and Po is obtained. This yields the coordinates of any lane line discrete point Po under the reference line coordinates, which is also the first intermediate non-reference line lane line data RLD2.

[0128] Step S402: Along the s-axis of the reference line coordinate system, the lane lines corresponding to the first intermediate non-reference line lane line data are evenly divided into multiple segments. Each segment contains multiple lane line discrete points under the reference coordinate system to obtain the second intermediate non-reference line lane line data.

[0129] This embodiment mainly involves segmenting the first intermediate non-reference line lane line data RLD2 to obtain the second intermediate non-reference line lane line data RLD3.

[0130] Step S403: Apply the least squares method to the discrete points of the lane lines in the reference coordinate system in each segment of the second intermediate non-reference line lane line data, and take points at uniform intervals along the s-axis of the reference line coordinate system to obtain the third intermediate non-reference line lane line data.

[0131] The third middle non-reference lane line data is represented as RLD4.

[0132] Step S404: Based on the definition of lane numbering in OpenDRIVE, determine the lane number corresponding to each lane line in the third intermediate non-reference line lane line data to obtain the fourth intermediate non-reference line lane line data.

[0133] Specifically, based on the lane numbering definition in OpenDRIVE, when determining the lane number corresponding to the third middle non-reference lane line data RLD4, the middle lane has no width and serves as a reference for lane numbering. The lane number of the middle lane is 0, and the other lane numbers start from the middle lane, decreasing to the right (representing the negative t-axis of the reference line coordinate system) and increasing to the left (representing the positive t-axis of the reference line coordinate system), thus obtaining the fourth middle non-reference lane line data RLD5.

[0134] Step S405: Based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane line data, determine the lane width corresponding to each lane line discrete point to obtain the fifth intermediate non-reference line lane line data.

[0135] Specifically, for lane line R1 with lane number less than 0, its adjacent lane line R2 in the positive t-axis direction can be found. For any discrete point P_R1 of R1, there is a corresponding P_R2 in L2 with the same s-coordinate. The absolute value of the difference between the t-coordinates of P_R1 and P_R2 is the lane width corresponding to lane line R1. For lane line L1 with lane number greater than 0, its adjacent lane line L2 in the negative t-axis direction can be found. For any discrete point P_L1 of L1, there is a corresponding P_L2 in L2 with the same s-coordinate. The absolute value of the difference between the t-coordinates of P_L1 and P_L2 is the lane width corresponding to lane line L1. After determining the lane width corresponding to each discrete point of the lane line, the fifth intermediate non-reference lane line data RLD6 is obtained.

[0136] Step S406: When the number of lanes changes each time, based on the definition of lane section in OpenDRIVE, the lane section is re-divided for the fifth middle non-reference line lane line data to obtain the sixth middle non-reference line lane line data.

[0137] Specifically, according to the definition of a lane section in OpenDRIVE, each lane section contains a fixed number of lanes. Each time the number of lanes changes, a new lane section is needed to describe it. All lane line discrete points are traversed in ascending order along the road reference line's s-coordinate. If the OpenDRIVE lane number to which a lane line discrete point belongs changes between two s-coordinate values, a lane section is divided. The larger s-value is used as the starting s-coordinate of the newly appearing lane section, and the OpenDRIVE lane numbers and lane discrete points contained in the previous lane section are counted, thus obtaining the sixth intermediate non-reference line lane line data RLD7.

[0138] Step S407: Use the least squares method to piecewise fit the lane width in the sixth intermediate non-reference line lane line data to obtain the target lane data.

[0139] Specifically, the lane width at a point in OpenDRIVE is described by a cubic polynomial of the distance difference between that point and the starting point of the lane section along the reference line. For each OpenDRIVE lane in each lane section, it is uniformly divided into several small segments, and the offset_start of the starting position of each segment relative to the starting position of the lane section in the reference line coordinate system is recorded. For the discrete points of the lane line contained in each segment, the difference in the s value between each discrete point and the starting point of each segment in the reference line coordinate system is obtained. Then, the cubic polynomial expression of the lane width of each segment is fitted to obtain the target lane data RLD8.

[0140] To further optimize the above embodiments, see [link to relevant documentation]. Figure 5 The flowchart of a method for obtaining road height data disclosed in this embodiment of the invention, specifically step S106, may include:

[0141] Step S501: Based on the target reference line lane line data, convert the vehicle's inertial navigation position information corresponding to all timestamps in the vehicle's inertial navigation data in the global coordinate system to the reference line coordinate system to obtain the initial road height data;

[0142] Specifically, in OpenDRIVE, road height is described along a reference line. Based on the target reference line lane line data CLD5, the vehicle's inertial navigation position information in the global coordinate system of all timestamps in the vehicle's inertial navigation data VD2 is transformed to the reference line coordinate system, obtaining the discrete s and t coordinates of the vehicle's inertial navigation position in the reference line coordinate system. The z-axis height information of each point in the global coordinate system is retained, resulting in the initial road height data ED1.

[0143] Among them, the inertial navigation data VD2 of the vehicle in the global coordinate system is obtained by transforming the actual inertial navigation data VD1 from the latitude and longitude coordinate system to the global coordinate system.

[0144] Step S502: Use the least squares method to piecewise fit the road height data of the initial road height data to obtain the target road height data.

[0145] Specifically, the system is uniformly divided into several small segments along the s-axis of the reference line coordinate system, and the s-offset of the starting point of each segment is recorded. In each segment, the s-discrete value ds between the discrete point of the vehicle's inertial navigation position and the starting point of the segment is calculated from the s-coordinate value of the discrete point of the vehicle's inertial navigation position. The cubic polynomial expression of the lane height of each segment is fitted with the ds value of the discrete point of the vehicle's inertial navigation position and the height value to obtain the target road height data ED2.

[0146] To further optimize the above embodiments, see [link to relevant documentation]. Figure 6 This invention discloses a method for generating road map-related parameters from target reference line lane line data, target lane data, and target road height data according to the OpenDRIVE protocol, and constructing an OpenDRIVE road map. The method includes:

[0147] Step S601: Generate reference line descriptions in the OpenDRIVE road map from the target reference line lane line data according to the OpenDRIVE protocol;

[0148] Specifically, in OpenDRIVE, the reference line is described using the geometry element within the planView element of the road. In the target reference line lane line data CLD5, each curve segment corresponds to a geometry; the s-coordinate of the curve's starting point is described by the s attribute of the geometry element; the coordinates of the curve's starting point in the global coordinate system are described by the x and y attributes of the geometry element; the angle of the tangent vector at the curve's starting point is described by the hdg attribute of the geometry element; the curve length is described by the length attribute of the geometry element; and the coefficients of the parametric equations in the local coordinate system are described using paramPoly3 in the geometry element, corresponding to the aU, bU, cU, dU, aV, bV, cV, and dV attributes.

[0149] The formulas involved in the existing OpenDRIVE specification are as follows:

[0150] u(p)=aU+bU*p+cU*p2+dU*p 3

[0151] v(p)=aV+bV*p+cV*p²+dV*p 3

[0152] Where u and v correspond to the u-axis and v-axis of the local coordinate system; p is an intermediate parameter; and aU, bU, cU, dU, aV, bV, cV, and dV are coefficients.

[0153] Step S602: Generate road descriptions in the OpenDRIVE road map from the target lane data RLD8 according to the OpenDRIVE protocol;

[0154] Specifically, in OpenDRIVE, lanes are described within the `lanes` element of the `road` element. In the target lane data RLD8, each lane section corresponds to a `laneSection` description; the 's' coordinate of the starting position of each lane section corresponds to the 's' attribute in the `laneSection`; each `laneSection` is divided into three parts: left, center, and right. Lanes with lane numbers greater than 0 are described in the `lane` element of the left part, lanes with lane numbers less than 0 are described in the `lane` element of the right part, and lane 0 is described in the center part, which defines the driving direction of the road, has no width attribute, and is equivalent to the reference line. In the lane data RLD8, the lane width attribute of each segment within each lane is described in the `width` element of the `lane` element. The 's' offset of the starting position of each segment relative to the starting position of its lane section corresponds to the `sOffset` attribute in the `width` element, and the coefficients of the cubic polynomial of the fitted lane width correspond to the `a`, `b`, `c`, and `d` attributes in the `width` element.

[0155] In this context, the a, b, c, and d attributes of the width element are also coefficients of a cubic polynomial, expressed as follows:

[0156] Width(ds) = a + b * ds + c * ds 2 +d*ds 3 ;

[0157] Where ds is the distance along the reference line between the given position and the starting point of the segment; Width is the lane width at the given position; and a, b, c, and d are coefficients.

[0158] For an explanation of how the Width(ds) expression works, please refer to existing mature technologies; it will not be elaborated here.

[0159] Step S603: Generate a road height description in the OpenDRIVE road map based on the target road height data according to the OpenDRIVE protocol.

[0160] Specifically, in OpenDRIVE, road height is described in the `elevation` element within the `elevationProfile` element. In the target road height data ED2, each segment corresponds to one elevation; the starting point `s` bias of each segment corresponds to the `s` attribute in the elevation; and the cubic polynomial coefficients of the fitted road height correspond to the `a`, `b`, `c`, and `d` attributes in the elevation element.

[0161] The cubic polynomial expression for the fitted road height is as follows:

[0162] elev(ds) = a + b * ds + c * ds 2 +d*ds 3 ;

[0163] Where ds is the distance along the reference line between the given position and the starting point of the segment; elev is the road height at the given position; and a, b, c, and d are coefficients.

[0164] For the working principle of the cubic polynomial elev(ds), please refer to existing mature solutions; it will not be elaborated here.

[0165] Corresponding to the above method embodiments, the present invention also discloses a road map construction apparatus.

[0166] See Figure 7 The present invention discloses a schematic diagram of a road map construction device, which includes:

[0167] Acquisition unit 701 is used to acquire real lane line data and real inertial navigation data in the vehicle coordinate system;

[0168] In this invention, real lane line data and real inertial navigation data are acquired by a road survey vehicle, which refers to a vehicle that collects data on real roads. In this embodiment, the road survey vehicle is equipped with a visual sensor (e.g., a camera) and an inertial navigation sensor. The visual sensor collects real lane line data, and the inertial navigation sensor collects real inertial navigation data.

[0169] Real-world lane line data includes time-stamped lane line IDs and cubic polynomial coefficients of the lane lines in the vehicle coordinate system after inverse perspective transformation. In practical applications, in addition to collecting real-world lane line data, the vision sensor can also collect time-stamped traffic sign and traffic light information. Traffic light information includes: traffic light ID, traffic light type, and traffic light location.

[0170] The actual inertial navigation data, also known as the vehicle's inertial navigation pose data, includes: latitude and longitude with timestamps, altitude, vehicle yaw angle, roll angle, and pitch angle information.

[0171] The lane line data determination unit 702 is used to determine the discrete point information of each lane line in the global coordinate system based on the real lane line data and the real inertial navigation data, so as to obtain the target lane line data.

[0172] Lane line data segmentation unit 703 is used to select a global lane line from the target lane line data as a reference lane line, and divide the target lane line data into reference lane line data and non-reference lane line data based on the reference lane line.

[0173] The first data processing unit 704 is used to process the reference line lane line data using a preset reference line processing scheme to obtain the target reference line lane line data.

[0174] The second data processing unit 705 is used to process the non-reference line lane line data using a preset non-reference line processing scheme to obtain the target lane data.

[0175] The third data processing unit 706 is used to process the vehicle's inertial navigation data in the global coordinate system based on the target reference line lane line data to obtain the target road height data. The vehicle's inertial navigation data is obtained by converting the real inertial navigation data from the latitude and longitude coordinate system to the global coordinate system.

[0176] Map building unit 707 is used to generate road map-related parameters by taking target reference line lane line data, target lane data and target road height data according to the OpenDRIVE protocol, and build an OpenDRIVE road map.

[0177] The road map parameters include: reference line description, lane description, and road height description.

[0178] In summary, this invention discloses a road map construction device that acquires real lane line data and real inertial navigation data in a vehicle coordinate system. Based on the real lane line data and real inertial navigation data, it determines the discrete point information of each lane line belonging to each lane line in the global coordinate system to obtain target lane line data. From the target lane line data, it selects a global lane line as a reference lane line. Based on the reference lane line, it divides the target lane line data into reference lane line data and non-reference lane line data. It processes the reference lane line data using a preset reference line processing scheme to obtain target reference lane line data, and processes the non-reference lane line data using a preset non-reference line processing scheme to obtain target lane data. Based on the target reference lane line data, it processes the vehicle's inertial navigation data in the global coordinate system to obtain target road height data. It then generates road map-related parameters according to the OpenDRIVE protocol using the target reference lane line data, target lane data, and target road height data to construct an OpenDRIVE road map. As can be seen, this invention uses real lane line data and real inertial navigation data from visual perception to build an OpenDRIVE road map. Since the visual perception results of each frame are based on the vehicle's own coordinate system, the constructed OpenDRIVE road map has high reliability and can accurately test the performance of intelligent driving algorithms under real road maps. It is applicable to fields such as intelligent driving simulation testing and can greatly improve the accuracy of intelligent driving simulation test results.

[0179] To further optimize the above embodiments, the lane line data determination unit 702 specifically includes:

[0180] The discrete point extraction subunit is used to extract a preset number of discrete points at uniform intervals from the lane curve corresponding to the real lane line data to obtain the first lane line data.

[0181] The first coordinate system transformation subunit is used to convert the discrete point coordinates of the lane line in the vehicle coordinate system into the discrete point coordinates of the lane line in the global coordinate system based on the first lane line data and the global pose data of the vehicle coordinate origin, so as to obtain the second lane line data.

[0182] The clustering subunit is used to cluster the second lane line data using a clustering algorithm to obtain the target lane line data.

[0183] To further optimize the above embodiments, the first data processing unit 704 specifically includes:

[0184] The first segmentation subunit is used to divide the reference lane line corresponding to the reference lane line data into multiple segments at uniform intervals. Each segment contains multiple lane line discrete points in the global coordinate system to obtain the first intermediate reference lane line data.

[0185] The first fitting subunit is used to fit a straight line to the discrete points of the lane lines in each segment of the first intermediate reference lane line data using the least squares method, and select the midpoint of the straight line to obtain the second intermediate reference lane line data.

[0186] The second fitting subunit is used to fit the discrete points of the lane lines in the second intermediate reference lane line data three times to obtain the B-spline curve, and to determine the curve parameters of the B-spline curve to obtain the third intermediate reference lane line data.

[0187] The second coordinate system transformation subunit is used to transform the parametric equations of each curve segment in the third intermediate reference lane line data from the global coordinate system to the local coordinate system to obtain the target reference lane line data.

[0188] To further optimize the above embodiments, the second data processing unit 705 specifically includes:

[0189] The third coordinate system transformation subunit is used to transform all lane line discrete points in the non-reference lane line data from the global coordinate system to the reference coordinate system to obtain the first intermediate non-reference lane line data.

[0190] The second segmentation subunit is used to divide the lane lines corresponding to the first intermediate non-reference line lane line data into multiple segments at uniform intervals along the s-axis direction of the reference line coordinate system. Each segment contains multiple lane line discrete points under the reference coordinate system to obtain the second intermediate non-reference line lane line data.

[0191] The third fitting subunit is used to perform least squares linear fitting on the discrete points of the lane lines in the reference coordinate system in each segment of the second intermediate non-reference line lane line data, and to take points at uniform intervals along the s-axis of the reference line coordinate system to obtain the third intermediate non-reference line lane line data.

[0192] The lane number determination subunit is used to determine the lane number corresponding to each lane line in the third intermediate non-reference line lane line data based on the definition of lane number in OpenDRIVE, so as to obtain the fourth intermediate non-reference line lane line data.

[0193] The lane width determination subunit is used to determine the lane width corresponding to each lane line discrete point based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane line data, and to obtain the fifth intermediate non-reference line lane line data.

[0194] The lane line data partitioning subunit is used to re-divide the lane line data of the fifth intermediate non-reference line based on the definition of lane section in OpenDRIVE each time the number of lanes changes, to obtain the lane line data of the sixth intermediate non-reference line.

[0195] The fourth fitting subunit is used to fit the lane width in the sixth intermediate non-reference line lane line data piecewise using the least squares method to obtain the target lane data.

[0196] To further optimize the above embodiments, the third data processing unit specifically includes:

[0197] The fourth coordinate system transformation subunit is used to transform the vehicle's inertial navigation position information corresponding to all timestamps in the global coordinate system to the reference line coordinate system based on the target reference line lane line data, so as to obtain the initial road height data.

[0198] The fifth fitting subunit is used to fit the road height piecewise using the least squares method on the initial road height data to obtain the target road height data.

[0199] To further optimize the above embodiments, the map building unit specifically includes:

[0200] The reference line description generation sub-unit is used to generate reference line descriptions in the OpenDRIVE road map from the target reference line lane line data according to the OpenDRIVE protocol.

[0201] The road description generation subunit is used to generate road descriptions in the OpenDRIVE road map from the target lane data RLD8 according to the OpenDRIVE protocol.

[0202] The road height description generation sub-unit is used to generate road height descriptions in the OpenDRIVE road map from the target road height data according to the OpenDRIVE protocol.

[0203] It should be noted that for the specific working principles of each component in the device embodiment, please refer to the corresponding section of the method embodiment, which will not be repeated here.

[0204] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0205] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0206] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for constructing a road map, characterized in that, include: Acquire real lane line data and real inertial navigation data in the vehicle coordinate system; Based on the real lane line data and the real inertial navigation data, determine the discrete point information of each lane line in the global coordinate system to obtain the target lane line data; Select a global lane line from the target lane line data as a reference lane line, and divide the target lane line data into reference lane line data and non-reference lane line data based on the reference lane line. The reference line lane line data is processed using a preset reference line processing scheme to obtain the target reference line lane line data; The non-reference line lane data is processed using a preset non-reference line processing scheme to obtain the target lane data; Based on the target reference line lane line data, the vehicle's inertial navigation data in the global coordinate system is processed to obtain the target road height data. The vehicle's inertial navigation data is obtained by converting the actual inertial navigation data from the latitude and longitude coordinate system to the global coordinate system. The target reference line lane line data, the target lane data, and the target road height data are used to generate road map-related parameters according to the OpenDRIVE protocol to construct an OpenDRIVE road map.

2. The construction method according to claim 1, characterized in that, The process of determining the discrete point information of each lane line belonging to each lane line in the global coordinate system based on the real lane line data and the real inertial navigation data to obtain the target lane line data includes: A preset number of discrete points are extracted at uniform intervals from the lane curves corresponding to the real lane line data to obtain the first lane line data; Based on the first lane line data and the global pose data of the vehicle coordinate origin, the discrete point coordinates of the lane line in the vehicle coordinate system are converted into the discrete point coordinates of the lane line in the global coordinate system to obtain the second lane line data. The target lane line data is obtained by clustering the second lane line data using a clustering algorithm.

3. The construction method according to claim 1, characterized in that, The process of processing the reference line lane line data using a preset reference line processing scheme to obtain the target reference line lane line data includes: The reference lane line data corresponding to the reference lane line data is evenly divided into multiple segments, each segment containing multiple lane line discrete points in the global coordinate system, to obtain the first intermediate reference lane line data. For each segment of the first intermediate reference lane line data, the lane line discrete points are fitted with a straight line using the least squares method, and the midpoint of the straight line is selected to obtain the second intermediate reference lane line data. The lane line discrete points in the second intermediate reference lane line data are piecewise fitted three times to obtain a B-spline curve, and the curve parameters of the B-spline curve are determined to obtain the third intermediate reference lane line data. The parametric equations of each curve segment in the third intermediate reference lane line data under the global coordinate system are transformed to the local coordinate system to obtain the target reference line lane line data.

4. The construction method according to claim 1, characterized in that, The process of processing the non-reference line lane data using a preset non-reference line processing scheme to obtain target lane data includes: Transform all lane line discrete points in the non-reference lane line data from the global coordinate system to the reference coordinate system to obtain the first intermediate non-reference lane line data. Along the s-axis of the reference coordinate system, the lane lines corresponding to the first intermediate non-reference line lane line data are evenly divided into multiple segments, each of which contains multiple lane line discrete points in the reference coordinate system, to obtain the second intermediate non-reference line lane line data. The least squares method is used to linearly fit the discrete points of the lane lines in the reference coordinate system in each segment of the second intermediate non-reference line lane line data, and points are taken at uniform intervals along the s-axis of the reference coordinate system to obtain the third intermediate non-reference line lane line data. Based on the definition of lane numbering in OpenDRIVE, the lane number corresponding to each lane line in the third intermediate non-reference line lane line data is determined to obtain the fourth intermediate non-reference line lane line data. Based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane line data, the lane width corresponding to each lane line discrete point is determined to obtain the fifth intermediate non-reference line lane line data. When the number of lanes changes each time, based on the definition of lane section in OpenDRIVE, the lane section is re-divided for the fifth intermediate non-reference lane line data to obtain the sixth intermediate non-reference lane line data. The target lane data is obtained by piecewise fitting of the lane width in the sixth intermediate non-reference line lane line data using the least squares method.

5. The construction method according to claim 1, characterized in that, The process of processing the vehicle's inertial navigation data in the global coordinate system based on the target reference line lane line data to obtain target road height data includes: Based on the target reference line lane line data, the vehicle inertial navigation position information corresponding to all timestamps in the vehicle inertial navigation data in the global coordinate system is transformed to the reference coordinate system to obtain the initial road height data; The initial road height data is piecewise fitted using the least squares method to obtain the target road height data.

6. The construction method according to claim 1, characterized in that, The step of generating road map-related parameters from the target reference line lane line data, the target lane data, and the target road height data according to the OpenDRIVE protocol to construct an OpenDRIVE road map includes: The target reference line lane line data is used to generate reference line descriptions in the OpenDRIVE road map according to the OpenDRIVE protocol; The target lane data RLD8 is used to generate road descriptions in the OpenDRIVE road map according to the OpenDRIVE protocol; The target road height data is used to generate a road height description in the OpenDRIVE road map according to the OpenDRIVE protocol.

7. A road map construction apparatus, characterized in that, include: The acquisition unit is used to acquire real lane line data and real inertial navigation data in the vehicle coordinate system; The lane line data determination unit is used to determine the discrete point information of each lane line belonging to each lane line in the global coordinate system based on the real lane line data and the real inertial navigation data, so as to obtain the target lane line data. The lane line data segmentation unit is used to select a global lane line from the target lane line data as a reference lane line, and divide the target lane line data into reference lane line data and non-reference lane line data based on the reference lane line. The first data processing unit is used to process the reference line lane line data using a preset reference line processing scheme to obtain the target reference line lane line data. The second data processing unit is used to process the non-reference line lane line data using a preset non-reference line processing scheme to obtain target lane data. The third data processing unit is used to process the vehicle's inertial navigation data in the global coordinate system based on the target reference line lane line data to obtain target road height data, wherein the vehicle's inertial navigation data is obtained by converting the real inertial navigation data from the latitude and longitude coordinate system to the global coordinate system; The map building unit is used to generate road map-related parameters from the target reference line lane line data, the target lane data, and the target road height data according to the OpenDRIVE protocol, and to build an OpenDRIVE road map.

8. The construction apparatus according to claim 7, characterized in that, The lane line data determination unit specifically includes: The discrete point extraction subunit is used to extract a preset number of discrete points at uniform intervals from the lane curve corresponding to the real lane line data to obtain the first lane line data. The first coordinate system transformation subunit is used to convert the discrete point coordinates of the lane line in the vehicle coordinate system into the discrete point coordinates of the lane line in the global coordinate system based on the first lane line data and the global pose data of the vehicle coordinate origin, so as to obtain the second lane line data. The clustering subunit is used to cluster the second lane line data using a clustering algorithm to obtain the target lane line data.

9. The construction apparatus according to claim 7, characterized in that, The first data processing unit specifically includes: The first segmentation subunit is used to divide the reference lane line corresponding to the reference lane line data into multiple segments at uniform intervals. Each segment contains multiple lane line discrete points in the global coordinate system to obtain the first intermediate reference lane line data. The first fitting subunit is used to fit a straight line to the discrete points of the lane lines in each segment of the first intermediate reference lane line data using the least squares method, and select the midpoint of the straight line to obtain the second intermediate reference lane line data. The second fitting subunit is used to fit the discrete points of the lane lines in the second intermediate reference lane line data three times to obtain the B-spline curve, and to determine the curve parameters of the B-spline curve to obtain the third intermediate reference lane line data. The second coordinate system transformation subunit is used to transform the parametric equations of each curve segment in the third intermediate reference lane line data from the global coordinate system to the local coordinate system to obtain the target reference line lane line data.

10. The construction apparatus according to claim 7, characterized in that, The second data processing unit specifically includes: The third coordinate system transformation subunit is used to transform all lane line discrete points in the non-reference lane line data from the global coordinate system to the reference coordinate system to obtain the first intermediate non-reference lane line data. The second segmentation subunit is used to divide the lane lines corresponding to the first intermediate non-reference line lane line data into multiple segments at uniform intervals along the s-axis direction of the reference coordinate system. Each segment contains multiple lane line discrete points under the reference coordinate system to obtain the second intermediate non-reference line lane line data. The third fitting subunit is used to perform least squares linear fitting on the discrete points of the lane lines in the reference coordinate system in each segment of the second intermediate non-reference line lane line data, and to take points at uniform intervals along the s-axis of the reference coordinate system to obtain the third intermediate non-reference line lane line data. The lane number determination subunit is used to determine the lane number corresponding to each lane line in the third intermediate non-reference line lane line data based on the definition of lane number in OpenDRIVE, so as to obtain the fourth intermediate non-reference line lane line data. The lane width determination subunit is used to determine the lane width corresponding to each lane line discrete point based on the lane number corresponding to each lane line in the fourth intermediate non-reference line lane line data, and to obtain the fifth intermediate non-reference line lane line data. The lane line data division subunit is used to re-divide the lane section of the fifth intermediate non-reference lane line data based on the definition of lane section in OpenDRIVE each time the number of lanes changes, so as to obtain the sixth intermediate non-reference lane line data. The fourth fitting subunit is used to fit the lane width in the sixth intermediate non-reference line lane line data piecewise using the least squares method to obtain the target lane data.

Images