WO2021016803A1 - 高精度地图定位方法、系统、平台及计算机可读存储介质 - Google Patents

高精度地图定位方法、系统、平台及计算机可读存储介质 Download PDF

Info

Publication number
WO2021016803A1
WO2021016803A1 PCT/CN2019/098172 CN2019098172W WO2021016803A1 WO 2021016803 A1 WO2021016803 A1 WO 2021016803A1 CN 2019098172 W CN2019098172 W CN 2019098172W WO 2021016803 A1 WO2021016803 A1 WO 2021016803A1
Authority
WO
WIPO (PCT)
Prior art keywords
map
driving environment
point cloud
movable platform
grid
Prior art date
Application number
PCT/CN2019/098172
Other languages
English (en)
French (fr)
Inventor
江灿森
钟阳
孙路
周游
Original Assignee
深圳市大疆创新科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳市大疆创新科技有限公司 filed Critical 深圳市大疆创新科技有限公司
Priority to PCT/CN2019/098172 priority Critical patent/WO2021016803A1/zh
Priority to CN201980033698.8A priority patent/CN112154303B/zh
Publication of WO2021016803A1 publication Critical patent/WO2021016803A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
    • G01C21/28Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
    • G01C21/30Map- or contour-matching
    • G01C21/32Structuring or formatting of map data
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01CMEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
    • G01C21/00Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
    • G01C21/26Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network

Definitions

  • This application relates to the technical field of high-precision maps, and in particular to a high-precision map positioning method, system, platform and computer-readable storage medium.
  • a high-resolution map is also called a high-definition map (HD Map), which is a map that can serve autonomous driving.
  • HD Map high-definition map
  • Collect the original point cloud data of the area traveled by the movable platform process the collected original point cloud data through high-precision inertial navigation and point cloud registration technology, register and optimize the pose, and obtain high-precision three-dimensional point cloud data , And generate a corresponding high-precision map based on the registered three-dimensional point cloud data.
  • a large-scale high-precision map is generated based on the registered three-dimensional point cloud data. It is usually used that all point clouds are calculated and generated based on the height of the gravity direction, and the subsequent use of high-precision map positioning is also performed according to the gravity direction; however, Since there are often uphills, downhills, etc. in actual driving scenes, this method of generating high-precision maps according to the direction of gravity and using high-precision maps for positioning does not handle the scenes of driving on non-horizontal surfaces well. It may lead to poor positioning accuracy of the movable platform in this driving scenario, which may affect subsequent decisions of the movable platform. Therefore, how to improve the positioning accuracy of the movable platform is a problem to be solved urgently.
  • this application provides a high-precision map positioning method, system, platform, and computer-readable storage medium, aiming to improve the positioning accuracy of a mobile platform.
  • this application provides a high-precision map positioning method, including:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the present application also provides a driving system, the driving system includes a lidar, a memory, and a processor; the memory is used to store a computer program;
  • the processor is configured to execute the computer program, and when executing the computer program, implement the following steps:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the present application also provides a movable platform, the movable platform includes a lidar, a memory, and a processor; the memory is used to store a computer program;
  • the processor is configured to execute the computer program, and when executing the computer program, implement the following steps:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor realizes:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the embodiments of the present application provide a high-precision map positioning method, system, platform, and computer-readable storage medium.
  • determining the current driving environment where the mobile platform is located and obtaining the point cloud matching rule corresponding to the current driving environment, based on The point cloud matching rule matches the online point cloud map with the offline high-precision map to obtain the map matching result.
  • the positioning result of the movable platform is determined based on the map matching result, and the adaptive point is based on the current driving environment where the movable platform is located.
  • Cloud matching rules can improve the positioning accuracy of the mobile platform in various driving environments and effectively improve the positioning accuracy of the mobile platform.
  • FIG. 1 is a schematic flowchart of the steps of a high-precision map positioning method provided by an embodiment of the present application
  • FIG. 2 is a schematic flowchart of sub-steps of the high-precision map positioning method in FIG. 1;
  • FIG. 3 is a schematic flowchart of steps of another high-precision map positioning method provided by an embodiment of the present application.
  • FIG. 4 is a schematic block diagram of the structure of a driving system provided by an embodiment of the present application.
  • FIG. 5 is a schematic block diagram of the structure of a movable platform provided by an embodiment of the present application.
  • FIG. 1 is a schematic flowchart of steps of a high-precision map positioning method provided by an embodiment of the present application.
  • the high-precision map positioning method can be applied to a movable platform or a driving system.
  • movable platforms include vehicles and aircraft.
  • Aircraft include unmanned aerial vehicles and manned aerial vehicles.
  • Vehicles include manned and unmanned vehicles.
  • Unmanned aerial vehicles include rotary-wing unmanned aerial vehicles, such as four-rotor unmanned aerial vehicles and six-rotor unmanned aerial vehicles.
  • Human aircraft, eight-rotor unmanned aerial vehicles, fixed-wing unmanned aerial vehicles, or a combination of rotor-type and fixed-wing unmanned aerial vehicles, are not limited here.
  • the high-precision map positioning method includes step S101 to step S104.
  • the mobile platform uses high-precision lidar to collect three-dimensional point cloud data from the traveled area, and uses high-precision inertial navigation system and point cloud registration algorithm to process the collected three-dimensional point cloud data to generate offline high Accuracy map.
  • use high-precision lidar to collect three-dimensional point cloud data from the traveled area use the Inertial Measurement Unit (IMU) to collect the attitude data of the movable platform, and use the Global Positioning System (GPS) to collect Based on the position data of the movable platform, the collected 3D point cloud data is fused based on the attitude data and the position data, and an offline high-precision map is generated based on the fused 3D point cloud data.
  • IMU Inertial Measurement Unit
  • GPS Global Positioning System
  • the mobile platform acquires offline high-precision maps during the movement, and collects 3D point cloud data of objects around the mobile platform in real time through lidar, and establishes an online point cloud map based on the real-time collected 3D point cloud data.
  • the lidar can determine the three-dimensional point cloud data of the object based on the distance between the laser emission point and the reflection point of the emitted laser light on the object, and the emission direction of the laser at the laser emission point.
  • the three-dimensional point cloud data of objects around the movable platform includes the distance between the object and the movable platform, the angle between the object and the movable platform, and the three-dimensional coordinates of the object.
  • the three-dimensional point cloud collected by the movable platform is projected in the direction of gravity to form a two-dimensional point cloud map.
  • This projection method is relatively simple, but in some non-horizontal plane points There may be errors in cloud processing.
  • the projection method used in this application can solve this problem.
  • the offline high-precision map generation can no longer be generated by using the gravity direction projection method.
  • the movable platform divides the pre-collected three-dimensional point cloud data into three-dimensional point cloud data corresponding to a first preset driving environment and three-dimensional point cloud data corresponding to a second preset driving environment; based on the road normal vector projection rule, Project the three-dimensional point cloud data corresponding to the first preset driving environment to a two-dimensional grid to generate a high-precision map corresponding to the first preset driving environment; based on the road normal vector projection rule, the second preset driving environment corresponds to The three-dimensional point cloud data is projected to a two-dimensional grid to generate a high-precision map corresponding to the second preset driving environment; the high-precision map corresponding to the first preset driving environment and the high-precision map corresponding to the second preset driving environment are performed Fusion, get offline high-precision map.
  • the normal vector projection of the road surface can better preserve the three-dimensional results of local positions, and can more accurately express the geometric characteristics of the scene.
  • accurate offline high-precision maps can be obtained, which is convenient for subsequent Carry out positioning in the driving environment to improve the accuracy of positioning.
  • the offline high-precision map positioning can also no longer use the gravity direction projection method.
  • the processing of the three-dimensional point cloud data collected by the movable platform in real time can also be divided into a first preset driving environment and a second preset driving environment according to the environment where the movable platform is located, and both are based on the road normal vector projection rule To process the real-time collected three-dimensional point cloud data, so as to achieve subsequent high-precision map positioning.
  • the first preset driving environment is a non-level road driving environment
  • the second preset environment is a horizontal road driving environment.
  • the method for generating the high-precision map corresponding to the first preset driving environment is specifically: converting the three-dimensional point cloud data corresponding to the first preset driving environment into the first three-dimensional point cloud data in the global map coordinate system; The first three-dimensional point cloud data in the coordinate system is converted into the second three-dimensional point cloud data in the local map coordinate system corresponding to the first preset driving environment; the second three-dimensional point cloud data is normalized according to the vertical axis direction of the local map coordinate system And project it to a two-dimensional grid to generate a high-precision map corresponding to the first preset driving environment.
  • the three-dimensional point cloud data corresponding to the second preset driving environment is data in the inertial coordinate system
  • the projection direction of the road surface normal vector is aligned with the vertical axis of the inertial coordinate system
  • the high-precision map corresponding to the second preset driving environment is generated
  • the specific method is: orthogonally projecting the three-dimensional point cloud data corresponding to the second preset driving environment to a two-dimensional grid according to the vertical axis of the inertial coordinate system to generate a high-precision map corresponding to the second preset driving environment.
  • the 3D point cloud data in the first preset driving environment and the second preset driving environment are related to their specific driving environment.
  • the projection direction of the 3D point cloud data is different in the non-level road driving environment and the non-level road driving environment.
  • S102 Determine the current driving environment in which the mobile platform is located, and obtain a point cloud matching rule corresponding to the current driving environment.
  • the mobile platform determines the current driving environment where the mobile platform is located, and obtains the point cloud matching rule corresponding to the current driving environment.
  • the current driving environment of the mobile platform includes the road surface direction
  • the point cloud matching rule includes the projection direction of the road surface normal vector.
  • the projection direction of the road surface normal vector changes with the road surface direction in the current driving environment, and the projection of the road surface normal vector The direction is perpendicular to the direction of the road, or the projection direction of the road normal vector is perpendicular to the average direction of the local road.
  • the projection direction of the road surface normal vector changes with the change of the road surface direction in the current driving environment, which can better save the three-dimensional results of the local position, can more accurately express the geometric characteristics of the scene, and improve the positioning accuracy of the local position.
  • the road surface is obtained by fitting the three-dimensional point cloud data corresponding to the current driving environment.
  • the road surface includes simple horizontal roads, simple non-horizontal roads, and composite roads composed of horizontal and non-horizontal roads. Including the horizontal road direction and the non-horizontal road direction.
  • the road direction of a purely horizontal road is the horizontal road direction
  • the road direction of a pure non-horizontal road is the non-horizontal road direction.
  • the horizontal road direction is perpendicular to the direction of gravity and is also in line with the normal vector
  • the projection direction is vertical
  • the non-horizontal road surface is not perpendicular to the direction of gravity, and perpendicular to the projection direction of the road normal vector.
  • the projection direction of the road normal vector is perpendicular to the average direction of the local road (non-horizontal road).
  • step S102 includes sub-steps S1021 to S1022.
  • the current geographic location of the movable platform is acquired, where the current geographic location is the geographic location output by the positioning system of the movable platform at the current moment, and the positioning system may be a global positioning system.
  • the current driving environment includes a horizontal road driving environment and a non-horizontal road driving environment.
  • the horizontal road driving environment includes a driving environment on a completely level road surface and a driving environment on an incomplete level road surface.
  • the movable platform determines the position coordinates of the movable platform in the offline high-precision map according to the current geographic location; obtains the position coordinate set corresponding to the non-level road driving environment, and determines whether the position coordinate is in the position coordinate set; If the coordinates are in the position coordinate set, it is determined that the current driving environment where the movable platform is located is a non-level road driving environment. It should be noted that the above-mentioned position coordinate set is set based on actual conditions, which is not specifically limited in this application.
  • the method for determining the location coordinates of the movable platform in the offline high-precision map is specifically: obtaining the location coordinates of the movable platform from the current location, marking the location coordinates in the offline high-precision map, and then obtaining The location coordinates of the surrounding objects in the offline high-precision map are determined by the location coordinates of the surrounding objects in the offline high-precision map, and the position coordinates of the movable platform in the offline high-precision map are determined.
  • the current geographic location of the mobile platform and the offline high-precision map the current driving environment where the mobile platform is located can be accurately determined.
  • the current driving environment can also be determined by: acquiring current attitude data of the movable platform, where the current attitude data is the attitude data output by the inertial measurement unit of the movable platform at the current moment; according to the current attitude data , To determine the current driving environment where the movable platform is located.
  • the pitch or roll tilt angle of the movable platform determines whether the pitch or roll tilt angle meets the preset tilt angle range, if the pitch and roll The tilt angle satisfies the preset tilt angle range, for example, the tilt degree is less than a certain threshold, the movable platform is determined to be in a horizontal road driving environment, if the pitch tilt angle or the roll tilt angle does not meet the preset tilt angle range, then Make sure that the movable platform is in a non-level road environment.
  • the above-mentioned tilt angle range can be set based on actual conditions, which is not specifically limited in this application.
  • the pitch angle or roll angle is the pitch angle or roll angle output by the inertial measurement unit mounted on the movable platform.
  • the current driving environment of the movable platform can be accurately determined through the current posture data of the movable platform.
  • the current driving environment can be determined by: acquiring the current environment image of the current driving environment where the movable platform is located, and generating a corresponding depth map according to the current environment image; determining the available depth map according to the generated depth map The current driving environment in which the mobile platform is located.
  • the movable platform is provided with a binocular stereo camera, the current environment image of the current driving environment where the movable platform is located can be captured by the binocular stereo camera, the optical axis of the binocular stereo camera and the chassis of the movable platform parallel. By combining the binocular stereo camera with the depth map, the current driving environment of the movable platform can be accurately determined.
  • the movable platform calculates the vertical distance between the road point and the optical axis of the binocular stereo camera based on the generated depth map, and if the vertical distance is greater than or less than the preset distance threshold, it is determined that the movable platform is in a non-level road driving environment If the vertical distance is equal to the preset distance threshold, it is determined that the movable platform is in a horizontal road driving environment.
  • the above-mentioned distance threshold can be set based on actual conditions, which is not specifically limited in this application.
  • the movable platform can also be based on at least two of the positioning system, binocular stereo camera and inertial measurement unit to determine the current driving environment where the movable platform is located, that is, it can be based on the positioning system and the binocular stereo camera.
  • Determine the current driving environment of the movable platform or determine the current driving environment of the movable platform based on the positioning system and inertial measurement unit, or determine the current driving environment of the movable platform based on the binocular stereo camera and the inertial measurement unit Driving environment, or based on the positioning system, binocular stereo camera and inertial measurement unit to determine the current driving environment where the movable platform is located.
  • the candidate positioning result set of the movable platform is determined, wherein the candidate positioning result set includes at least two candidate positioning results; based on the projection direction of the road surface normal vector in the point cloud matching rule, each candidate positioning result is projected to an online point Cloud map, obtain the online grid map corresponding to each candidate positioning result; calculate the loss cost value between the online grid map and offline high-precision map corresponding to each candidate positioning result; correspond each candidate positioning result to its own The cost value of the loss between the online raster map and the offline high-precision map is used as the map matching result.
  • the method for determining the candidate positioning result set can be: obtaining the current position data and current attitude data of the movable platform; determining the candidate position set according to the current position data and the preset position error value; according to the current attitude data and the preset attitude error value Determine the candidate pose set; determine the candidate positioning result set according to the candidate position set and the candidate pose set.
  • the current position data of the movable platform is the position data output by the positioning system of the movable platform at the current moment
  • the current posture data of the movable platform is the posture data output by the inertial measurement unit of the movable platform at the current moment.
  • the position data includes the geographic coordinates of the movable platform, and the attitude data includes the pitch angle, roll angle, and yaw angle of the movable platform. It should be noted that the above-mentioned preset position error value and preset attitude error value can be set based on actual conditions, which is not specifically limited in this application. Through the current position data and current posture data of the movable platform, the candidate positioning result set can be accurately determined, thereby indirectly improving the accuracy of positioning.
  • the calculation method of the loss cost value is specifically: rasterize the offline high-precision map to obtain the offline raster map corresponding to each online raster map, wherein the online raster map and the offline raster map correspond to each other The number of grids in the map is the same; according to the height of each grid in each offline grid map and the height of each grid in each online grid map, the loss between each online grid map and the offline high-precision map is calculated Cost value.
  • the height of each grid in the online grid map and the height of each grid in the offline grid map are both the average height of the point cloud in the grid.
  • the cost value of the loss between the online grid map and the offline high-precision map corresponding to each candidate positioning result is obtained from the map matching result; the candidate positioning result corresponding to the smallest loss cost value is used as the mobile platform Positioning results.
  • the candidate positioning result set includes candidate positioning result A, candidate positioning result B, and candidate positioning result C, and in the order of loss cost value, the resulting arrangement order is candidate positioning result B, candidate positioning result C, candidate positioning result A , That is, the candidate positioning result corresponding to the smallest loss cost value is the candidate positioning result A, then the candidate positioning result A is taken as the target positioning result.
  • the high-precision map positioning method determines the current driving environment where the mobile platform is located, and obtains the point cloud matching rule corresponding to the current driving environment, and then based on the point cloud matching rule, compares the online point cloud map with the offline The high-precision map is matched to obtain the map matching result. Finally, the positioning result of the movable platform is determined based on the map matching result. Based on the current driving environment where the movable platform is located, the adaptive point cloud matching rule can improve the mobility of the movable platform in various driving conditions. The positioning accuracy in the environment can effectively improve the positioning accuracy of the movable platform.
  • FIG. 3 is a schematic flowchart of the steps of another high-precision map positioning method according to an embodiment of the present application.
  • the high-precision map positioning method includes steps S201 to S209.
  • the movable platform acquires an offline high-precision map during its movement, collects 3D point cloud data of objects around the movable platform in real time through lidar, and establishes an online point cloud map based on the real-time collected 3D point cloud data.
  • S202 Determine the current driving environment where the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment.
  • the mobile platform determines the current driving environment where the mobile platform is located, and obtains the point cloud matching rule corresponding to the current driving environment.
  • the current driving environment of the mobile platform includes the road surface direction
  • the point cloud matching rule includes the projection direction of the road surface normal vector.
  • the projection direction of the road surface normal vector changes with the road surface direction in the current driving environment, and the projection of the road surface normal vector The direction is perpendicular to the direction of the road, or the projection direction of the road normal vector is perpendicular to the average direction of the local road.
  • S203 Determine a candidate positioning result set of the movable platform, wherein the candidate positioning result set includes at least two candidate positioning results.
  • the candidate positioning result set of the movable platform is determined, that is, the current position data and current attitude data of the movable platform are obtained; according to the current position data and preset position error
  • the candidate position set is determined by the value; the candidate pose set is determined according to the current pose data and the preset pose error value; the candidate positioning result set is determined according to the candidate position set and the candidate pose set.
  • the current position data of the movable platform is the position data output by the positioning system of the movable platform at the current moment
  • the current posture data of the movable platform is the posture data output by the inertial measurement unit of the movable platform at the current moment.
  • the position data includes the geographic coordinates of the movable platform, and the attitude data includes the pitch angle, roll angle, and yaw angle of the movable platform. It should be noted that the above-mentioned preset position error value and preset attitude error value can be set based on actual conditions, which is not specifically limited in this application.
  • the mobile platform obtains the driving environment label from the point cloud matching rule, and determines whether the driving environment label is a tunnel The label corresponding to the driving environment, if the driving environment label is a label corresponding to the tunnel driving environment, it can be determined that the current driving environment of the movable platform is the tunnel driving environment.
  • the mobile platform projects each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule, and it can be obtained that there is no mark High-level online grid map.
  • the movable platform calculates and marks the average height of each grid according to the height of each point that meets the preset conditions in each grid projected to obtain the corresponding marked height of each candidate positioning result. Online raster map.
  • the preset condition is that the height of the points in the grid is less than the preset threshold.
  • the preset threshold can be set based on actual conditions. This application does not specifically limit this. Optionally, the preset threshold is 3 meters.
  • the offline high-precision map is rasterized to obtain the offline raster map corresponding to each online raster map, wherein the number of grids of the online raster map and the offline raster map corresponding to each other are the same;
  • the height of each grid in each offline grid map and the height of each grid in each online grid map are calculated, and the cost value of the loss between each online grid map and the offline high-precision map is calculated.
  • the height of each grid in the online grid map and the height of each grid in the offline grid map are both the average height of the point cloud in the grid.
  • the mobile platform obtains the loss cost value between the online grid map and the offline high-precision map corresponding to each candidate positioning result from the map matching result; the candidate positioning result corresponding to the smallest loss cost value is used as the candidate positioning result.
  • the positioning result of the mobile platform is used as the candidate positioning result.
  • the height difference between adjacent grids in the horizontal direction can be increased, the horizontal deviation caused by the horizontal positioning can be reduced, and the mobility can be improved.
  • the positioning accuracy and accuracy of the platform in the tunnel scene can be improved.
  • FIG. 4 is a schematic block diagram of a driving system provided by an embodiment of the present application.
  • the driving system includes an unmanned driving system and a manned driving system.
  • the driving system 300 includes a processor 301, a memory 302, and a lidar 303, and the processor 301, the memory 302, and the lidar 303 are connected by a bus 304, such as an I2C (Inter-integrated Circuit) bus.
  • I2C Inter-integrated Circuit
  • the processor 301 may be a micro-controller unit (MCU), a central processing unit (CPU), a digital signal processor (Digital Signal Processor, DSP), or the like.
  • MCU micro-controller unit
  • CPU central processing unit
  • DSP Digital Signal Processor
  • the memory 302 may be a Flash chip, a read-only memory (ROM, Read-Only Memory) disk, an optical disk, a U disk, or a mobile hard disk.
  • the processor 301 and the memory 302 are the computing platform of the driving system
  • the lidar 303 may be an external device of the driving system or an internal component of the driving system, which is not specifically limited in this application.
  • the processor 301 is configured to run a computer program stored in the memory 302, and implement the following steps when executing the computer program:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the processor realizes the determination of the current driving environment in which the movable platform is located, it is used to realize:
  • the current driving environment where the movable platform is located is determined.
  • the driving system further includes a positioning system, and the current geographic position is a geographic position output by the positioning system at the current moment.
  • the processor when the processor is used to determine the current driving environment in which the mobile platform is located according to the current geographic location and the offline high-precision map, it is used to achieve:
  • the position coordinates are in the position coordinate set, it is determined that the current driving environment where the movable platform is located is a non-level road driving environment.
  • the driving system further includes an inertial measurement unit, and the processor is used to realize: when determining the current driving environment in which the movable platform is located:
  • the current driving environment where the movable platform is located is determined.
  • the processor realizes the determination of the current driving environment in which the movable platform is located, it is used to realize:
  • the current driving environment where the movable platform is located is determined.
  • the processor when the processor implements matching the online point cloud map with the offline high-precision map according to the point cloud matching rule to obtain a map matching result, it is used to implement:
  • the cost value of the loss between the online grid map and the offline high-precision map corresponding to each of the candidate positioning results is used as a map matching result.
  • the processor realizes the determination of the positioning result of the movable platform according to the map matching result, it is configured to realize:
  • the candidate positioning result corresponding to the smallest lost cost value is used as the positioning result of the movable platform.
  • the processor realizes the calculation of the lost cost value between the online grid map and the offline high-precision map corresponding to each candidate positioning result, it is used to realize:
  • each grid in each offline grid map calculates the distance between each online grid map and the offline high-precision map Lost cost value.
  • the processor projects each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule to obtain each candidate positioning result
  • the processor projects each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule to obtain each candidate positioning result
  • the driving environment tag is a tag corresponding to the tunnel driving environment, project each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule;
  • the preset condition is that the height of the points in the grid is less than a preset threshold.
  • the processor is also used to achieve:
  • the high-precision map corresponding to the first preset driving environment and the high-precision map corresponding to the second preset driving environment are merged to obtain an offline high-precision map.
  • the processor implements projection of the three-dimensional point cloud data corresponding to the first preset driving environment to a two-dimensional grid based on the road surface normal vector projection rule, so as to generate the data corresponding to the first preset driving environment
  • the processor implements projection of the three-dimensional point cloud data corresponding to the first preset driving environment to a two-dimensional grid based on the road surface normal vector projection rule, so as to generate the data corresponding to the first preset driving environment
  • the three-dimensional point cloud data corresponding to the second preset driving environment is data in an inertial coordinate system, and the projection direction of the road surface normal vector is aligned with the vertical axis of the inertial coordinate system;
  • the normal vector projection rule when projecting the three-dimensional point cloud data corresponding to the second preset driving environment to a two-dimensional grid to generate a high-precision map corresponding to the second preset driving environment, is used to achieve:
  • FIG. 5 is a schematic block diagram of a movable platform provided by an embodiment of the present application.
  • the mobile platform 400 includes a processor 401, a memory 402, and a lidar 403.
  • the processor 401, the memory 402, and the lidar 403 are connected by a bus 404, which is, for example, an I2C (Inter-integrated Circuit) bus.
  • movable platforms include vehicles and aircraft.
  • Aircraft include unmanned aerial vehicles and manned aerial vehicles.
  • Vehicles include manned and unmanned vehicles.
  • Unmanned aerial vehicles include rotary-wing unmanned aerial vehicles, such as four-rotor unmanned aerial vehicles and hexarotors.
  • Unmanned aerial vehicles, eight-rotor unmanned aerial vehicles can also be fixed-wing unmanned aerial vehicles, or a combination of rotary-wing and fixed-wing unmanned aerial vehicles, which are not limited here.
  • the processor 401 may be a micro-controller unit (MCU), a central processing unit (CPU), a digital signal processor (Digital Signal Processor, DSP), or the like.
  • MCU micro-controller unit
  • CPU central processing unit
  • DSP Digital Signal Processor
  • the memory 402 may be a Flash chip, a read-only memory (ROM, Read-Only Memory) disk, an optical disk, a U disk, or a mobile hard disk.
  • the processor 401 and the memory 402 are the computing platform of the driving system
  • the lidar 403 may be an external device of the driving system or an internal component of the driving system, which is not specifically limited in this application.
  • the processor 401 is configured to run a computer program stored in the memory 402, and implement the following steps when executing the computer program:
  • the movable platform Determine the current driving environment in which the movable platform is located, and obtain a point cloud matching rule corresponding to the current driving environment, wherein the current driving environment includes a road surface direction, and the point cloud matching rule includes a projection direction of a road surface normal vector, The projection direction changes as the road surface direction changes;
  • the positioning result of the movable platform is determined according to the map matching result.
  • the processor realizes the determination of the current driving environment in which the movable platform is located, it is used to realize:
  • the current driving environment where the movable platform is located is determined.
  • the movable platform further includes a positioning system, and the current geographic location is a geographic location output by the positioning system at the current moment.
  • the processor when the processor is used to determine the current driving environment in which the mobile platform is located according to the current geographic location and the offline high-precision map, it is used to achieve:
  • the position coordinates are in the position coordinate set, it is determined that the current driving environment where the movable platform is located is a non-level road driving environment.
  • the movable platform further includes an inertial measurement unit, and the processor is used to achieve: when determining the current driving environment where the movable platform is located:
  • the current driving environment where the movable platform is located is determined.
  • the processor realizes the determination of the current driving environment in which the movable platform is located, it is used to realize:
  • the current driving environment where the movable platform is located is determined.
  • the processor when the processor implements matching the online point cloud map with the offline high-precision map according to the point cloud matching rule to obtain a map matching result, it is used to implement:
  • the cost value of the loss between the online grid map and the offline high-precision map corresponding to each of the candidate positioning results is used as a map matching result.
  • the processor realizes the determination of the positioning result of the movable platform according to the map matching result, it is configured to realize:
  • the candidate positioning result corresponding to the smallest lost cost value is used as the positioning result of the movable platform.
  • the processor realizes the calculation of the lost cost value between the online grid map and the offline high-precision map corresponding to each candidate positioning result, it is used to realize:
  • each grid in each offline grid map calculates the distance between each online grid map and the offline high-precision map Lost cost value.
  • the processor projects each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule to obtain each candidate positioning result
  • the processor projects each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule to obtain each candidate positioning result
  • the driving environment tag is a tag corresponding to the tunnel driving environment, project each candidate positioning result to the online point cloud map based on the projection direction of the road surface normal vector in the point cloud matching rule;
  • the preset condition is that the height of the points in the grid is less than a preset threshold.
  • the processor implements the acquisition of the offline high-precision map, it is also used to implement:
  • the high-precision map corresponding to the first preset driving environment and the high-precision map corresponding to the second preset driving environment are merged to obtain an offline high-precision map.
  • the processor implements projection of the three-dimensional point cloud data corresponding to the first preset driving environment to a two-dimensional grid based on the road surface normal vector projection rule, so as to generate the data corresponding to the first preset driving environment
  • the processor implements projection of the three-dimensional point cloud data corresponding to the first preset driving environment to a two-dimensional grid based on the road surface normal vector projection rule, so as to generate the data corresponding to the first preset driving environment
  • the three-dimensional point cloud data corresponding to the second preset driving environment is data in an inertial coordinate system, and the projection direction of the road surface normal vector is aligned with the vertical axis of the inertial coordinate system;
  • the normal vector projection rule when projecting the three-dimensional point cloud data corresponding to the second preset driving environment to a two-dimensional grid to generate a high-precision map corresponding to the second preset driving environment, is used to achieve:
  • the embodiments of the present application also provide a computer-readable storage medium, the computer-readable storage medium stores a computer program, the computer program includes program instructions, and the processor executes the program instructions to implement the foregoing implementation The steps of the high-precision map positioning method provided in the example.
  • the computer-readable storage medium may be the internal storage unit of the driving system or the movable platform according to any of the foregoing embodiments, for example, the hard disk or memory of the driving system or the movable platform.
  • the computer-readable storage medium may also be an external storage device of the driving system or a removable platform, such as a plug-in hard disk or a smart memory card (Smart Media Card, SMC) equipped on the driving system or the removable platform. , Secure Digital (SD) card, Flash Card (Flash Card), etc.
  • SD Secure Digital
  • Flash Card Flash Card

Abstract

本申请公开了一种高精度地图定位方法、系统、平台及计算机可读存储介质,该方法包括:获取离线高精度地图,并建立在线点云地图(S101);获取当前行驶环境对应的点云匹配规则(S102);根据点云匹配规则,将在线点云地图与离线高精度地图进行匹配(S103);根据匹配结果确定定位结果(S104)。本申请提高定位准确性。

Description

高精度地图定位方法、系统、平台及计算机可读存储介质 技术领域
本申请涉及高精度地图的技术领域,尤其涉及一种高精度地图定位方法、系统、平台及计算机可读存储介质。
背景技术
高精度地图也称为高分辨率地图(High Definition Map,HD Map),是一可以为自动驾驶服务的地图。采集可移动平台行驶过的区域的原始点云数据,通过高精度惯导和点云配准技术对采集到的原始点云数据进行处理,配准优化位姿,得到高精度的三维点云数据,并依据配准好的三维点云数据生成对应的高精度地图。
目前,依据配准好的三维点云数据生成大范围的高精度地图,通常采用的是所有点云均依据重力方向的高度进行计算生成,并且后续使用高精度地图定位也依据重力方向进行;然而由于实际驾驶场景中经常会有上坡、下坡等,这种按照重力方向的方式生成高精度地图并使用高精度地图定位的方式,并不能很好地处理在非水平面上驾驶的场景,从而可能导致可移动平台在这种驾驶场景下的定位精度不好而影响后续可移动平台的决策。因此,如何提高可移动平台的定位准确性是目前亟待解决的问题。
发明内容
基于此,本申请提供了一种高精度地图定位方法、系统、平台及计算机可读存储介质,旨在提高可移动平台的定位准确性。
第一方面,本申请提供了一种高精度地图定位方法,包括:
获取离线高精度地图,并建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
第二方面,本申请还提供了一种驾驶系统,所述驾驶系统包括激光雷达、存储器和处理器;所述存储器用于存储计算机程序;
所述处理器,用于执行所述计算机程序并在执行所述计算机程序时,实现如下步骤:
获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
第三方面,本申请还提供了一种可移动平台,所述可移动平台包括激光雷达、存储器和处理器;所述存储器用于存储计算机程序;
所述处理器,用于执行所述计算机程序并在执行所述计算机程序时,实现如下步骤:
获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
第四方面,本申请还提供了一种计算机可读存储介质,其特征在于,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行时使所述处理器实现:
获取离线高精度地图,并建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
本申请实施例提供了一种高精度地图定位方法、系统、平台及计算机可读存储介质,通过确定可移动平台所处的当前行驶环境,并获取当前行驶环境对应的点云匹配规则,然后基于该点云匹配规则,将在线点云地图与离线高精度地图进行匹配,得到地图匹配结果,最后基于地图匹配结果确定可移动平台的定位结果,基于可移动平台所处的当前行驶环境自适应点云匹配规则,可以提高可移动平台在各种行驶环境下的定位精度,有效的提高可移动平台的定位准确性。
应当理解的是,以上的一般描述和后文的细节描述仅是示例性和解释性的,并不能限制本申请。
附图说明
为了更清楚地说明本申请实施例技术方案,下面将对实施例描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1是本申请一实施例提供的一种高精度地图定位方法的步骤示意流程图;
图2是图1中的高精度地图定位方法的子步骤示意流程图;
图3是本申请一实施例提供的另一种高精度地图定位方法的步骤示意流程图;
图4是本申请一实施例提供的一种驾驶系统的结构示意性框图;
图5是本申请一实施例提供的一种可移动平台的结构示意性框图。
具体实施方式
下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都属于本申请保护的范围。
附图中所示的流程图仅是示例说明,不是必须包括所有的内容和操作/步骤,也不是必须按所描述的顺序执行。例如,有的操作/步骤还可以分解、组合或部分合并,因此实际执行的顺序有可能根据实际情况改变。
下面结合附图,对本申请的一些实施方式作详细说明。在不冲突的情况下,下述的实施例及实施例中的特征可以相互组合。
请参阅图1,图1是本申请一实施例提供的一种高精度地图定位方法的步骤示意流程图。该高精度地图定位方法可以应用在可移动平台或驾驶系统中。其中可移动平台包括车辆和飞行器,飞行器包括无人飞行器和有人飞行器,车辆包括有人驾驶车辆和无人驾驶车辆等,无人飞行器包括旋翼型无人飞行器,例如四旋翼无人飞行器、六旋翼无人飞行器、八旋翼无人飞行器,也可以是固定翼无人飞行器,还可以是旋翼型与固定翼无人飞行器的组合,在此不作限定。
具体地,如图1所示,该高精度地图定位方法包括步骤S101至步骤S104。
S101、获取离线高精度地图,并建立在线点云地图。
其中,可移动平台通过高精度的激光雷达对行驶过的区域采集三维点云数据,并通过高精度惯导系统和点云配准算法,对采集到的三维点云数据进行处理,生成离线高精度地图。或者,通过高精度的激光雷达对行驶过的区域采集三维点云数据,通过惯性测量单元(Inertial Measurement Unit,IMU)采集可移动平台的姿态数据,以及通过全球定位系统(Global PositioningSystem,GPS)采集可移动平台的位置数据,然后基于姿态数据和位置数据,对采集到的三维点云数据进行融合处理,并基于融合处理后的三维点云数据生成离线高精度地图。结合惯性测量单元、全球定位系统和三维点云数据可以生成准确的离线高精度地图。
可移动平台在移动过程中,获取离线高精度地图,并通过激光雷达实时采集可移动平台周围物体的三维点云数据,且基于实时采集到的三维点云数据建立在线点云地图。其中,激光雷达可以基于激光发射点与发射出的激光在物体上的反射点的距离,以及激光发射点的激光的发射方向,确定物体的三维点云 数据。可移动平台周边物体的三维点云数据包括物体与可移动平台的距离,物体与可移动平台的角度,以及物体的三维坐标等数据。
在传统的三维点云投影方法中,对可移动平台所采集的三维点云均是以重力方向进行投影而形成二维点云地图,这种投影方式较为简便,但在一些非水平平面的点云的处理上可能存在误差。本申请所采用的投影方式可以解决这种问题。在本申请的一实施例中,首先,离线高精度地图的生成即可以不再采用重力方向投影方式来生成。具体的,可移动平台将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数据和第二预设行驶环境对应的三维点云数据;基于路面法向量投影规则,将第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成第一预设行驶环境对应的高精度地图;基于路面法向量投影规则,将第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成第二预设行驶环境对应的高精度地图;对第一预设行驶环境对应的高精度地图和第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。路面法向量投影,能更好的保存局部位置的三维结果,能更准确的表达场景的几何特征,通过融合不同行驶环境下的高精度地图,可以得到准确的离线高精度地图,便于后续在不同行驶环境下进行定位,提高定位的准确性。其次,离线高精度地图的定位也可以不再采用重力方向投影方式来进行。具体的,对可移动平台实时采集的三维点云数据的处理,也可以根据可移动平台所处环境划分成第一预设行驶环境和第二预设行驶环境,并且均基于路面法向量投影规则来对实时采集的三维点云数据进行处理,从而实现后续的高精度地图的定位。
需要说明的是,第一预设行驶环境为非水平路面行驶环境,第二预设环境为水平路面行驶环境。其中,第一预设行驶环境对应的高精度地图的生成方式具体为:将第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;将全局地图坐标系下的第一三维点云数据转换为第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;按照局部地图坐标系的竖轴方向将第二三维点云数据正交投影至二维栅格,以生成第一预设行驶环境对应的高精度地图。
其中,第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影方向与惯性坐标系的竖轴对齐,第二预设行驶环境对应的高精度地图的生成方式具体为:按照惯性坐标系的竖轴将第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成第二预设行驶环境对应的高精度地图。第一预设行驶环境与第二预设行驶环境下的三维点云数据与其具体行驶环 境有关,在非水平路面行驶环境和非水平路面行驶环境下,三维点云数据投影方向不同。
S102、确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则。
具体地,可移动平台在建立在线点云地图之后,确定可移动平台所处的当前行驶环境,并获取当前行驶环境对应的点云匹配规则。其中,可移动平台的当前行驶环境包括路面方向,点云匹配规则包括路面法向量的投影方向,路面法向量的投影方向随着当前行驶环境中的路面方向变化而变化,且路面法向量的投影方向与路面方向垂直,或者路面法向量的投影方向与局部路面的平均方向垂直。路面法向量的投影方向随着当前行驶环境中的路面方向变化而变化,能更好的保存局部位置的三维结果,能更准确的表达场景的几何特征,可以提高局部位置的定位准确性。
需要说明的是,路面为通过当前行驶环境对应的三维点云数据拟合得到的,该路面包括单纯的水平路面、单纯的非水平路面以及由水平路面和非水平路面组成的复合路面,路面方向包括水平路面方向和非水平路面方向,单纯的水平路面的路面方向为水平路面方向,单纯的非水平路面的路面方向为非水平路面方向,水平路面方向与重力方向垂直,也与路面法向量的投影方向垂直,而非水平路面不与重力方向垂直,与路面法向量的投影方向垂直,在复合路面中,路面法向量的投影方向与局部路面(非水平路面)的平均方向垂直。
在一实施例中,如图2所示,步骤S102包括子步骤S1021至S1022。
S1021、获取可移动平台的当前地理位置。
具体地,获取可移动平台的当前地理位置,其中,当前地理位置为该可移动平台的定位系统在当前时刻输出的地理位置,定位系统可选为全球定位系统。
S1022、根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
其中,当前行驶环境包括水平路面行驶环境和非水平路面行驶环境,水平路面行驶环境包括完全水平路面的行驶环境,也包括不完全水平路面的行驶环境。
具体地,可移动平台根据当前地理位置,确定可移动平台在离线高精度地图中的位置坐标;获取非水平路面行驶环境对应的位置坐标集,并确定该位置 坐标是否位于位置坐标集;若位置坐标位于位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。需要说明的是,上述位置坐标集基于实际情况进行设置,本申请对此不作具体限定。
其中,可移动平台在离线高精度地图中的位置坐标的确定方式具体为:从当前地理位置中获取可移动平台的地理位置坐标,并在该离线高精度地图中标记该地理位置坐标,然后获取该地理位置坐标周围物体在离线高精度地图中的位置坐标,并基于周围物体在离线高精度地图中的位置坐标,确定可移动平台在离线高精度地图中的位置坐标。通过可移动平台的当前地理位置和离线高精度地图,可以准确的确定可移动平台所处的当前行驶环境。
在一实施例中,当前行驶环境的确定方式还可以为:获取可移动平台的当前姿态数据,其中,当前姿态数据为可移动平台的惯性测量单元在当前时刻输出的姿态数据;根据当前姿态数据,确定可移动平台所处的当前行驶环境。
具体地,基于当前姿态数据,确定可移动平台的俯仰倾斜角或横滚倾斜角,并判断该俯仰倾斜角或横滚倾斜角是否满足预设的倾斜角范围,如果该俯仰倾斜角和横滚倾斜角满足预设的倾斜角范围,例如其倾斜程度小于某一阈值,则确定可移动平台处于水平路面行驶环境,如果该俯仰倾斜角或横滚倾斜角不满足预设的倾斜角范围,则确定可移动平台处于非水平路面行驶环境。需要说明的是,上述倾斜角范围可基于实际情况进行设置,本申请对此不作具体限定。可选地,该俯仰倾斜角或横滚倾斜角为搭载于可移动平台的惯性测量单元输出的俯仰角或横滚角。通过可移动平台的当前姿态数据可准确的确定可移动平台所处的当前行驶环境。
在一实施例中,当前行驶环境的确定方式还可以为:获取可移动平台所处的当前行驶环境的当前环境图像,并根据当前环境图像生成对应的深度图;根据生成的深度图,确定可移动平台所处的当前行驶环境。其中,该可移动平台设置有双目立体摄像机,通过该双目立体摄像机可以拍摄得到可移动平台所处的当前行驶环境的当前环境图像,该双目立体摄像机的光轴与可移动平台的底盘平行。通过双目立体摄像机结合深度图,可以准确的确定可移动平台所处的当前行驶环境。
具体地,可移动平台基于生成的深度图计算路面点距离双目立体摄像机的光轴的垂直距离,如果该垂直距离大于或小于预设的距离阈值,则确定可移动平台处于非水平路面行驶环境,如果该垂直距离等于预设的距离阈值,则确定可移动平台处于水平路面行驶环境。需要说明的是,上述距离阈值可基于实际 情况进行设置,本申请对此不作具体限定。
需要说明的是,可移动平台还可以基于定位系统、双目立体摄像机和惯性测量单元中的至少两个,融合确定可移动平台所处的当前行驶环境,即可以基于定位系统和双目立体摄像机确定可移动平台所处的当前行驶环境,或者可以基于定位系统和惯性测量单元确定可移动平台所处的当前行驶环境,或者可以基于双目立体摄像机和惯性测量单元确定可移动平台所处的当前行驶环境,或者基于定位系统、双目立体摄像机和惯性测量单元确定可移动平台所处的当前行驶环境。通过结合多个传感器的数据确定可移动平台所处的当前行驶环境,可以进一步准确地确定可移动平台所处的当前行驶环境。
S103、根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果。
具体地,确定可移动平台的候选定位结果集,其中,候选定位结果集包括至少两个候选定位结果;基于点云匹配规则中路面法向量的投影方向,将每个候选定位结果投影至在线点云地图,得到每个候选定位结果各自对应的在线栅格地图;计算每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值;将每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值作为地图匹配结果。
其中,候选定位结果集的确定方式可以为:获取可移动平台的当前位置数据和当前姿态数据;根据当前位置数据和预设位置误差值确定候选位置集;根据当前姿态数据和预设姿态误差值确定候选姿态集;根据候选位置集和所述候选姿态集,确定候选定位结果集。其中,可移动平台的当前位置数据为可移动平台的定位系统在当前时刻输出的位置数据,可移动平台的当前姿态数据为可移动平台的惯性测量单元在当前时刻输出的姿态数据。该位置数据包括可移动平台的地理位置坐标,该姿态数据包括可移动平台的俯仰角、横滚角和偏航角。需要说明的是,上述预设位置误差值和预设姿态误差值可基于实际情况进行设置,本申请对此不作具体限定。通过可移动平台的当前位置数据和当前姿态数据,可以准确的确定候选定位结果集,从而间接的提高定位的准确性。
其中,损失代价值的计算方式具体为:对离线高精度地图进行栅格化处理,得到每个在线栅格地图各自对应的离线栅格地图,其中,互相对应的在线栅格地图和离线栅格地图的栅格数量相同;根据每个离线栅格地图中各栅格的高度和每个在线栅格地图中各栅格的高度,计算每个在线栅格地图与离线高精度地图之间的损失代价值。其中,在线栅格地图中各栅格的高度和离线栅格地图中 各栅格的高度均为栅格内的点云的平均高度。
S104、根据所述地图匹配结果确定所述可移动平台的定位结果。
具体地,从该地图匹配结果中获取每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值;将最小的损失代价值对应的候选定位结果作为可移动平台的定位结果。
例如,候选定位结果集包括候选定位结果A、候选定位结果B和候选定位结果C,且按照损失代价值的大小顺序,得到的排列顺序为候选定位结果B、候选定位结果C、候选定位结果A,即最小的损失代价值对应的候选定位结果为候选定位结果A,则将候选定位结果A作为目标定位结果。
上述实施例提供的高精度地图定位方法,通过确定可移动平台所处的当前行驶环境,并获取当前行驶环境对应的点云匹配规则,然后基于该点云匹配规则,将在线点云地图与离线高精度地图进行匹配,得到地图匹配结果,最后基于地图匹配结果确定可移动平台的定位结果,基于可移动平台所处的当前行驶环境自适应点云匹配规则,可以提高可移动平台在各种行驶环境下的定位精度,有效的提高可移动平台的定位准确性。
请参阅图3,图3是本申请一实施例提供的另一种高精度地图定位方法的步骤示意流程图。
具体地,如图3所示,该高精度地图定位方法包括步骤S201至S209。
S201、获取离线高精度地图,并建立在线点云地图。
具体地,可移动平台在移动过程中,获取离线高精度地图,并通过激光雷达实时采集可移动平台周围物体的三维点云数据,且基于实时采集到的三维点云数据建立在线点云地图。
S202、确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则。
具体地,可移动平台在建立在线点云地图之后,确定可移动平台所处的当前行驶环境,并获取当前行驶环境对应的点云匹配规则。其中,可移动平台的当前行驶环境包括路面方向,点云匹配规则包括路面法向量的投影方向,路面法向量的投影方向随着当前行驶环境中的路面方向变化而变化,且路面法向量的投影方向与路面方向垂直,或者路面法向量的投影方向与局部路面的平均方向垂直。
S203、确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果。
具体地,在获取到当前行驶环境对应的点云匹配规则时,确定可移动平台的候选定位结果集,即获取可移动平台的当前位置数据和当前姿态数据;根据当前位置数据和预设位置误差值确定候选位置集;根据当前姿态数据和预设姿态误差值确定候选姿态集;根据候选位置集和所述候选姿态集,确定候选定位结果集。其中,可移动平台的当前位置数据为可移动平台的定位系统在当前时刻输出的位置数据,可移动平台的当前姿态数据为可移动平台的惯性测量单元在当前时刻输出的姿态数据。该位置数据包括可移动平台的地理位置坐标,该姿态数据包括可移动平台的俯仰角、横滚角和偏航角。需要说明的是,上述预设位置误差值和预设姿态误差值可基于实际情况进行设置,本申请对此不作具体限定。
S204、从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签。
具体地,可移动平台在根据点云匹配规则,将在线点云地图与离线高精度地图进行匹配的过程中,从该点云匹配规则中获取行驶环境标签,并确定该行驶环境标签是否为隧道行驶环境对应的标签,如果该行驶环境标签为隧道行驶环境对应的标签,则可以确定可移动平台的当前行驶环境为隧道行驶环境。
S205、若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图。
具体地,当行驶环境标签为隧道行驶环境对应的标签,则可移动平台基于该点云匹配规则中路面法向量的投影方向,将每个候选定位结果投影至在线点云地图,可以得到没有标记高度的在线栅格地图。
S206、根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
具体地,可移动平台根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个候选定位结果各自对应的标记有高度的在线栅格地图。需要说明的是,预设条件为栅格中的点的高度小于预设 阈值,预设阈值可基于实际情况进行设置,本申请对此不作具体限定,可选地,预设阈值为3米。
S207、计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值。
具体地,对离线高精度地图进行栅格化处理,得到每个在线栅格地图各自对应的离线栅格地图,其中,互相对应的在线栅格地图和离线栅格地图的栅格数量相同;根据每个离线栅格地图中各栅格的高度和每个在线栅格地图中各栅格的高度,计算每个在线栅格地图与离线高精度地图之间的损失代价值。其中,在线栅格地图中各栅格的高度和离线栅格地图中各栅格的高度均为栅格内的点云的平均高度。
S208、将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
具体地,在计算得到每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值,将每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值作为地图匹配结果。
S209、根据所述地图匹配结果确定所述可移动平台的定位结果。
具体地,可移动平台从该地图匹配结果中获取每个候选定位结果各自对应的在线栅格地图与离线高精度地图之间的损失代价值;将最小的损失代价值对应的候选定位结果作为可移动平台的定位结果。
上述实施例提供的高精度地图定位方法,在隧道行驶场景下,通过过滤高度较高的点云,可以提高横向相邻栅格中的高度差异,减少横向定位产生的横向偏差,可以提高可移动平台在隧道场景下的定位精度和准确性。
请参阅图4,图4是本申请一实施例提供的驾驶系统的示意性框图。在一种实施方式中,该驾驶系统包括无人驾驶系统和有人驾驶系统。进一步地,该驾驶系统300包括处理器301、存储器302和激光雷达303,处理器301、存储器302和激光雷达303通过总线304连接,该总线304比如为I2C(Inter-integrated Circuit)总线。
具体地,处理器301可以是微控制单元(Micro-controller Unit,MCU)、中央处理单元(Central Processing Unit,CPU)或数字信号处理器(Digital Signal Processor,DSP)等。
具体地,存储器302可以是Flash芯片、只读存储器(ROM,Read-Only Memory)磁盘、光盘、U盘或移动硬盘等。
具体地,处理器301和存储器302为驾驶系统的计算平台,该激光雷达303可以为驾驶系统的外接设备,也可以为驾驶系统的内部组件,本申请对此不作具体限定。
其中,所述处理器301用于运行存储在存储器302中的计算机程序,并在执行所述计算机程序时实现如下步骤:
获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
可选地,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台的当前地理位置;
根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
可选地,所述驾驶系统还包括定位系统,所述当前地理位置为所述定位系统在当前时刻输出的地理位置。
可选地,所述处理器在实现根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境时,用于实现:
根据所述当前地理位置,确定所述可移动平台在所述离线高精度地图中的位置坐标;
获取非水平路面行驶环境对应的位置坐标集,并确定所述位置坐标是否位于所述位置坐标集;
若所述位置坐标位于所述位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。
可选地,所述驾驶系统还包括惯性测量单元,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台的当前姿态数据,其中,所述当前姿态数据为所述惯性测量单元在当前时刻输出的姿态数据;
根据所述当前姿态数据,确定可移动平台所处的当前行驶环境。
可选地,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台所处的当前行驶环境的当前环境图像,并根据所述当前环境图像生成对应的深度图;
根据生成的所述深度图,确定可移动平台所处的当前行驶环境。
可选地,所述处理器在实现根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果时,用于实现:
确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果;
基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图;
计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
可选地,所述处理器在实现根据所述地图匹配结果确定所述可移动平台的定位结果时,用于实现:
从所述地图匹配结果中获取每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
将最小的所述损失代价值对应的候选定位结果作为所述可移动平台的定位结果。
可选地,所述处理器在实现计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值时,用于实现:
对所述离线高精度地图进行栅格化处理,得到每个所述在线栅格地图各自对应的离线栅格地图,其中,互相对应的所述在线栅格地图和所述离线栅格地图的栅格数量相同;
根据每个所述离线栅格地图中各栅格的高度和每个所述在线栅格地图中各栅格的高度,计算每个所述在线栅格地图与所述离线高精度地图之间的损失代 价值。
可选地,所述处理器在实现基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图时,用于实现:
从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签;
若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图;
根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
可选地,所述处理器实现获取离线高精度地图之前,还用于实现:
将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数据和第二预设行驶环境对应的三维点云数据;
基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图;
基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图;
对所述第一预设行驶环境对应的高精度地图和所述第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。
可选地,所述处理器实现基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图时,用于实现:
将所述第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;
将所述全局地图坐标系下的第一三维点云数据转换为所述第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;
按照所述局部地图坐标系的竖轴方向将所述第二三维点云数据正交投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图。
可选地,所述第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影方向与所述惯性坐标系的竖轴对齐;所述处理器实现基 于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图时,用于实现:
按照所述惯性坐标系的竖轴将所述第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图。
需要说明的是,所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的驾驶系统的工作过程,可以参考前述高精度地图定位方法实施例中的对应过程,在此不再赘述。
请参阅图5,图5是本申请一实施例提供的可移动平台的示意性框图。该可移动平台400包括处理器401、存储器402和激光雷达403,处理器401、存储器402和激光雷达403通过总线404连接,该总线404比如为I2C(Inter-integrated Circuit)总线。其中,可移动平台包括车辆和飞行器,飞行器包括无人飞行器和有人飞行器,车辆包括有人驾驶车辆和无人驾驶车辆等,无人飞行器包括旋翼型无人飞行器,例如四旋翼无人飞行器、六旋翼无人飞行器、八旋翼无人飞行器,也可以是固定翼无人飞行器,还可以是旋翼型与固定翼无人飞行器的组合,在此不作限定。
具体地,处理器401可以是微控制单元(Micro-controller Unit,MCU)、中央处理单元(Central Processing Unit,CPU)或数字信号处理器(Digital Signal Processor,DSP)等。
具体地,存储器402可以是Flash芯片、只读存储器(ROM,Read-Only Memory)磁盘、光盘、U盘或移动硬盘等。
具体地,处理器401和存储器402为驾驶系统的计算平台,该激光雷达403可以为驾驶系统的外接设备,也可以为驾驶系统的内部组件,本申请对此不作具体限定。
其中,所述处理器401用于运行存储在存储器402中的计算机程序,并在执行所述计算机程序时实现如下步骤:
获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
根据所述地图匹配结果确定所述可移动平台的定位结果。
可选地,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台的当前地理位置;
根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
可选地,所述可移动平台还包括定位系统,所述当前地理位置为所述定位系统在当前时刻输出的地理位置。
可选地,所述处理器在实现根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境时,用于实现:
根据所述当前地理位置,确定所述可移动平台在所述离线高精度地图中的位置坐标;
获取非水平路面行驶环境对应的位置坐标集,并确定所述位置坐标是否位于所述位置坐标集;
若所述位置坐标位于所述位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。
可选地,所述可移动平台还包括惯性测量单元,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台的当前姿态数据,其中,所述当前姿态数据为所述惯性测量单元在当前时刻输出的姿态数据;
根据所述当前姿态数据,确定可移动平台所处的当前行驶环境。
可选地,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
获取可移动平台所处的当前行驶环境的当前环境图像,并根据所述当前环境图像生成对应的深度图;
根据生成的所述深度图,确定可移动平台所处的当前行驶环境。
可选地,所述处理器在实现根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果时,用于实现:
确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果;
基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图;
计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
可选地,所述处理器在实现根据所述地图匹配结果确定所述可移动平台的定位结果时,用于实现:
从所述地图匹配结果中获取每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
将最小的所述损失代价值对应的候选定位结果作为所述可移动平台的定位结果。
可选地,所述处理器在实现计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值时,用于实现:
对所述离线高精度地图进行栅格化处理,得到每个所述在线栅格地图各自对应的离线栅格地图,其中,互相对应的所述在线栅格地图和所述离线栅格地图的栅格数量相同;
根据每个所述离线栅格地图中各栅格的高度和每个所述在线栅格地图中各栅格的高度,计算每个所述在线栅格地图与所述离线高精度地图之间的损失代价值。
可选地,所述处理器在实现基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图时,用于实现:
从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签;
若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图;
根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
可选地,所述处理器实现获取离线高精度地图之前,还用于实现:
将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数据和第二预设行驶环境对应的三维点云数据;
基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图;
基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图;
对所述第一预设行驶环境对应的高精度地图和所述第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。
可选地,所述处理器实现基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图时,用于实现:
将所述第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;
将所述全局地图坐标系下的第一三维点云数据转换为所述第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;
按照所述局部地图坐标系的竖轴方向将所述第二三维点云数据正交投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图。
可选地,所述第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影方向与所述惯性坐标系的竖轴对齐;所述处理器实现基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图时,用于实现:
按照所述惯性坐标系的竖轴将所述第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图。
需要说明的是,所属领域的技术人员可以清楚地了解到,为了描述的方便和简洁,上述描述的驾驶系统的工作过程,可以参考前述高精度地图定位方法实施例中的对应过程,在此不再赘述。
本申请的实施例中还提供一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序中包括程序指令,所述处理器执行所述程序指令,实现上述实施例提供的高精度地图定位方法的步骤。
其中,所述计算机可读存储介质可以是前述任一实施例所述的驾驶系统或 可移动平台的内部存储单元,例如所述驾驶系统或可移动平台的硬盘或内存。所述计算机可读存储介质也可以是所述驾驶系统或可移动平台的外部存储设备,例如所述驾驶系统或可移动平台上配备的插接式硬盘,智能存储卡(Smart Media Card,SMC),安全数字(Secure Digital,SD)卡,闪存卡(Flash Card)等。
应当理解,在此本申请说明书中所使用的术语仅仅是出于描述特定实施例的目的而并不意在限制本申请。如在本申请说明书和所附权利要求书中所使用的那样,除非上下文清楚地指明其它情况,否则单数形式的“一”、“一个”及“该”意在包括复数形式。
还应当理解,在本申请说明书和所附权利要求书中使用的术语“和/或”是指相关联列出的项中的一个或多个的任何组合以及所有可能组合,并且包括这些组合。
以上所述,仅为本申请的具体实施方式,但本申请的保护范围并不局限于此,任何熟悉本技术领域的技术人员在本申请揭露的技术范围内,可轻易想到各种等效的修改或替换,这些修改或替换都应涵盖在本申请的保护范围之内。因此,本申请的保护范围应以权利要求的保护范围为准。

Claims (40)

  1. 一种高精度地图定位方法,其特征在于,包括:
    获取离线高精度地图,并建立在线点云地图;
    确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
    根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
    根据所述地图匹配结果确定所述可移动平台的定位结果。
  2. 根据权利要求1所述的高精度地图定位方法,其特征在于,所述确定可移动平台所处的当前行驶环境,包括:
    获取可移动平台的当前地理位置;
    根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
  3. 根据权利要求2所述的高精度地图定位方法,其特征在于,所述当前地理位置为所述可移动平台的定位系统在当前时刻输出的地理位置。
  4. 根据权利要求2所述的高精度地图定位方法,其特征在于,所述根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境,包括:
    根据所述当前地理位置,确定所述可移动平台在所述离线高精度地图中的位置坐标;
    获取非水平路面行驶环境对应的位置坐标集,并确定所述位置坐标是否位于所述位置坐标集;
    若所述位置坐标位于所述位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。
  5. 根据权利要求1所述的高精度地图定位方法,其特征在于,所述确定可移动平台所处的当前行驶环境,包括:
    获取可移动平台的当前姿态数据,其中,所述当前姿态数据为所述可移动平台的惯性测量单元在当前时刻输出的姿态数据;
    根据所述当前姿态数据,确定可移动平台所处的当前行驶环境。
  6. 根据权利要求1所述的高精度地图定位方法,其特征在于,所述确定可 移动平台所处的当前行驶环境,包括:
    获取可移动平台所处的当前行驶环境的当前环境图像,并根据所述当前环境图像生成对应的深度图;
    根据生成的所述深度图,确定可移动平台所处的当前行驶环境。
  7. 根据权利要求1-6中任一项所述的高精度地图定位方法,其特征在于,所述根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果,包括:
    确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果;
    基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图;
    计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
  8. 根据权利要求7所述的高精度地图定位方法,其特征在于,所述根据所述地图匹配结果确定所述可移动平台的定位结果,包括:
    从所述地图匹配结果中获取每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将最小的所述损失代价值对应的候选定位结果作为所述可移动平台的定位结果。
  9. 根据权利要求8所述的高精度地图定位方法,其特征在于,所述计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值,包括:
    对所述离线高精度地图进行栅格化处理,得到每个所述在线栅格地图各自对应的离线栅格地图,其中,互相对应的所述在线栅格地图和所述离线栅格地图的栅格数量相同;
    根据每个所述离线栅格地图中各栅格的高度和每个所述在线栅格地图中各栅格的高度,计算每个所述在线栅格地图与所述离线高精度地图之间的损失代价值。
  10. 根据权利要求7所述的高精度地图定位方法,其特征在于,所述基于 所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图,包括:
    从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签;
    若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图;
    根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
  11. 根据权利要求1-6中任一项所述的高精度地图定位方法,其特征在于,所述获取离线高精度地图之前,还包括:
    将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数据和第二预设行驶环境对应的三维点云数据;
    基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图;
    基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图;
    对所述第一预设行驶环境对应的高精度地图和所述第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。
  12. 根据权利要求11所述的高精度地图定位方法,其特征在于,所述基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图,包括:
    将所述第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;
    将所述全局地图坐标系下的第一三维点云数据转换为所述第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;
    按照所述局部地图坐标系的竖轴方向将所述第二三维点云数据正交投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图。
  13. 根据权利要求11所述的高精度地图定位方法,其特征在于,所述第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影 方向与所述惯性坐标系的竖轴对齐;所述基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图,包括:
    按照所述惯性坐标系的竖轴将所述第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图。
  14. 一种驾驶系统,其特征在于,所述驾驶系统包括激光雷达、存储器和处理器;
    所述存储器用于存储计算机程序;
    所述处理器,用于执行所述计算机程序并在执行所述计算机程序时,实现如下步骤:
    获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
    确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
    根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
    根据所述地图匹配结果确定所述可移动平台的定位结果。
  15. 根据权利要求14所述的驾驶系统,其特征在于,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台的当前地理位置;
    根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
  16. 根据权利要求15所述的驾驶系统,其特征在于,所述驾驶系统还包括定位系统,所述当前地理位置为所述定位系统在当前时刻输出的地理位置。
  17. 根据权利要求15所述的驾驶系统,其特征在于,所述处理器在实现根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境时,用于实现:
    根据所述当前地理位置,确定所述可移动平台在所述离线高精度地图中的位置坐标;
    获取非水平路面行驶环境对应的位置坐标集,并确定所述位置坐标是否位于所述位置坐标集;
    若所述位置坐标位于所述位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。
  18. 根据权利要求14所述的驾驶系统,其特征在于,所述驾驶系统还包括惯性测量单元,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台的当前姿态数据,其中,所述当前姿态数据为所述惯性测量单元在当前时刻输出的姿态数据;
    根据所述当前姿态数据,确定可移动平台所处的当前行驶环境。
  19. 根据权利要求14所述的驾驶系统,其特征在于,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台所处的当前行驶环境的当前环境图像,并根据所述当前环境图像生成对应的深度图;
    根据生成的所述深度图,确定可移动平台所处的当前行驶环境。
  20. 根据权利要求14-19中任一项所述的驾驶系统,其特征在于,所述处理器在实现根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果时,用于实现:
    确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果;
    基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图;
    计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
  21. 根据权利要求20所述的驾驶系统,其特征在于,所述处理器在实现根据所述地图匹配结果确定所述可移动平台的定位结果时,用于实现:
    从所述地图匹配结果中获取每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将最小的所述损失代价值对应的候选定位结果作为所述可移动平台的定位结果。
  22. 根据权利要求21所述的驾驶系统,其特征在于,所述处理器在实现计 算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值时,用于实现:
    对所述离线高精度地图进行栅格化处理,得到每个所述在线栅格地图各自对应的离线栅格地图,其中,互相对应的所述在线栅格地图和所述离线栅格地图的栅格数量相同;
    根据每个所述离线栅格地图中各栅格的高度和每个所述在线栅格地图中各栅格的高度,计算每个所述在线栅格地图与所述离线高精度地图之间的损失代价值。
  23. 根据权利要求20所述的驾驶系统,其特征在于,所述处理器在实现基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图时,用于实现:
    从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签;
    若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图;
    根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
  24. 根据权利要求14-19中任一项所述的驾驶系统,其特征在于,所述处理器实现获取离线高精度地图之前,还用于实现:
    将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数据和第二预设行驶环境对应的三维点云数据;
    基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图;
    基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图;
    对所述第一预设行驶环境对应的高精度地图和所述第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。
  25. 根据权利要求24所述的驾驶系统,其特征在于,所述处理器实现基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二 维栅格,以生成所述第一预设行驶环境对应的高精度地图时,用于实现:
    将所述第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;
    将所述全局地图坐标系下的第一三维点云数据转换为所述第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;
    按照所述局部地图坐标系的竖轴方向将所述第二三维点云数据正交投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图。
  26. 根据权利要求24所述的驾驶系统,其特征在于,所述第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影方向与所述惯性坐标系的竖轴对齐;所述处理器实现基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图时,用于实现:
    按照所述惯性坐标系的竖轴将所述第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图。
  27. 一种可移动平台,其特征在于,所述可移动平台包括激光雷达、存储器和处理器;
    所述存储器用于存储计算机程序;
    所述处理器,用于执行所述计算机程序并在执行所述计算机程序时,实现如下步骤:
    获取离线高精度地图,并通过所述激光雷达采集到的三维点云数据建立在线点云地图;
    确定可移动平台所处的当前行驶环境,并获取所述当前行驶环境对应的点云匹配规则,其中,所述当前行驶环境包括路面方向,所述点云匹配规则包括路面法向量的投影方向,所述投影方向随着所述路面方向变化而变化;
    根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果;
    根据所述地图匹配结果确定所述可移动平台的定位结果。
  28. 根据权利要求27所述的可移动平台,其特征在于,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台的当前地理位置;
    根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境。
  29. 根据权利要求28所述的可移动平台,其特征在于,所述可移动平台还包括定位系统,所述当前地理位置为所述定位系统在当前时刻输出的地理位置。
  30. 根据权利要求28所述的可移动平台,其特征在于,所述处理器在实现根据所述当前地理位置和所述离线高精度地图,确定可移动平台所处的当前行驶环境时,用于实现:
    根据所述当前地理位置,确定所述可移动平台在所述离线高精度地图中的位置坐标;
    获取非水平路面行驶环境对应的位置坐标集,并确定所述位置坐标是否位于所述位置坐标集;
    若所述位置坐标位于所述位置坐标集,则确定可移动平台所处的当前行驶环境为非水平路面行驶环境。
  31. 根据权利要求27所述的可移动平台,其特征在于,所述可移动平台还包括惯性测量单元,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台的当前姿态数据,其中,所述当前姿态数据为所述惯性测量单元在当前时刻输出的姿态数据;
    根据所述当前姿态数据,确定可移动平台所处的当前行驶环境。
  32. 根据权利要求27所述的可移动平台,其特征在于,所述处理器在实现确定可移动平台所处的当前行驶环境时,用于实现:
    获取可移动平台所处的当前行驶环境的当前环境图像,并根据所述当前环境图像生成对应的深度图;
    根据生成的所述深度图,确定可移动平台所处的当前行驶环境。
  33. 根据权利要求27-32中任一项所述的可移动平台,其特征在于,所述处理器在实现根据所述点云匹配规则,将所述在线点云地图与所述离线高精度地图进行匹配,得到地图匹配结果时,用于实现:
    确定所述可移动平台的候选定位结果集,其中所述候选定位结果集包括至少两个候选定位结果;
    基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图;
    计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值作为地图匹配结果。
  34. 根据权利要求33所述的可移动平台,其特征在于,所述处理器在实现根据所述地图匹配结果确定所述可移动平台的定位结果时,用于实现:
    从所述地图匹配结果中获取每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值;
    将最小的所述损失代价值对应的候选定位结果作为所述可移动平台的定位结果。
  35. 根据权利要求34所述的可移动平台,其特征在于,所述处理器在实现计算每个所述候选定位结果各自对应的在线栅格地图与所述离线高精度地图之间的损失代价值时,用于实现:
    对所述离线高精度地图进行栅格化处理,得到每个所述在线栅格地图各自对应的离线栅格地图,其中,互相对应的所述在线栅格地图和所述离线栅格地图的栅格数量相同;
    根据每个所述离线栅格地图中各栅格的高度和每个所述在线栅格地图中各栅格的高度,计算每个所述在线栅格地图与所述离线高精度地图之间的损失代价值。
  36. 根据权利要求33所述的可移动平台,其特征在于,所述处理器在实现基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图,得到每个所述候选定位结果各自对应的在线栅格地图时,用于实现:
    从所述点云匹配规则中获取行驶环境标签,并确定所述行驶环境标签是否为隧道行驶环境对应的标签;
    若所述行驶环境标签为隧道行驶环境对应的标签,则基于所述点云匹配规则中路面法向量的投影方向,将每个所述候选定位结果投影至所述在线点云地图;
    根据投影到的每个栅格中满足预设条件的各点的高度,计算每个栅格的平均高度并标记,以得到每个所述候选定位结果各自对应的标记有高度的在线栅格地图,其中,所述预设条件为栅格中的点的高度小于预设阈值。
  37. 根据权利要求27-32中任一项所述的可移动平台,其特征在于,所述处理器实现获取离线高精度地图之前,还用于实现:
    将预先采集到的三维点云数据划分为第一预设行驶环境对应的三维点云数 据和第二预设行驶环境对应的三维点云数据;
    基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图;
    基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图;
    对所述第一预设行驶环境对应的高精度地图和所述第二预设行驶环境对应的高精度地图进行融合,得到离线高精度地图。
  38. 根据权利要求37所述的可移动平台,其特征在于,所述处理器实现基于路面法向量投影规则,将所述第一预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图时,用于实现:
    将所述第一预设行驶环境对应的三维点云数据转换为全局地图坐标系下的第一三维点云数据;
    将所述全局地图坐标系下的第一三维点云数据转换为所述第一预设行驶环境对应的局部地图坐标系下的第二三维点云数据;
    按照所述局部地图坐标系的竖轴方向将所述第二三维点云数据正交投影至二维栅格,以生成所述第一预设行驶环境对应的高精度地图。
  39. 根据权利要求37所述的可移动平台,其特征在于,所述第二预设行驶环境对应的三维点云数据为惯性坐标系下的数据,路面法向量的投影方向与所述惯性坐标系的竖轴对齐;所述处理器实现基于路面法向量投影规则,将所述第二预设行驶环境对应的三维点云数据投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图时,用于实现:
    按照所述惯性坐标系的竖轴将所述第二预设行驶环境对应的三维点云数据正交投影至二维栅格,以生成所述第二预设行驶环境对应的高精度地图。
  40. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行时使所述处理器实现如权利要求1至13中任一项所述的高精度地图定位方法。
PCT/CN2019/098172 2019-07-29 2019-07-29 高精度地图定位方法、系统、平台及计算机可读存储介质 WO2021016803A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2019/098172 WO2021016803A1 (zh) 2019-07-29 2019-07-29 高精度地图定位方法、系统、平台及计算机可读存储介质
CN201980033698.8A CN112154303B (zh) 2019-07-29 2019-07-29 高精度地图定位方法、系统、平台及计算机可读存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2019/098172 WO2021016803A1 (zh) 2019-07-29 2019-07-29 高精度地图定位方法、系统、平台及计算机可读存储介质

Publications (1)

Publication Number Publication Date
WO2021016803A1 true WO2021016803A1 (zh) 2021-02-04

Family

ID=73891471

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/098172 WO2021016803A1 (zh) 2019-07-29 2019-07-29 高精度地图定位方法、系统、平台及计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN112154303B (zh)
WO (1) WO2021016803A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112835370B (zh) * 2021-01-19 2023-07-14 北京小马智行科技有限公司 车辆的定位方法、装置、计算机可读存储介质及处理器
CN112950708B (zh) * 2021-02-05 2023-12-15 深圳市优必选科技股份有限公司 一种定位方法、定位装置及机器人
CN113607166B (zh) * 2021-10-08 2022-01-07 广东省科学院智能制造研究所 基于多传感融合的自主移动机器人室内外定位方法及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108228798A (zh) * 2017-12-29 2018-06-29 百度在线网络技术(北京)有限公司 确定点云数据之间的匹配关系的方法和装置
WO2018126083A1 (en) * 2016-12-30 2018-07-05 DeepMap Inc. Alignment of data captured by autonomous vehicle to generate high definition maps
CN108345008A (zh) * 2017-01-23 2018-07-31 郑州宇通客车股份有限公司 一种目标物检测方法、点云数据提取方法及装置
CN109443369A (zh) * 2018-08-20 2019-03-08 北京主线科技有限公司 利用激光雷达和视觉传感器融合构建动静态栅格地图的方法
CN109814572A (zh) * 2019-02-20 2019-05-28 广州市山丘智能科技有限公司 移动机器人定位建图方法、装置、移动机器人和存储介质

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106940186B (zh) * 2017-02-16 2019-09-24 华中科技大学 一种机器人自主定位与导航方法及系统
CN108732603B (zh) * 2017-04-17 2020-07-10 百度在线网络技术(北京)有限公司 用于定位车辆的方法和装置
CN108007453A (zh) * 2017-12-11 2018-05-08 北京奇虎科技有限公司 基于点云的地图更新方法、装置及电子设备
CN108303721B (zh) * 2018-02-12 2020-04-03 北京经纬恒润科技有限公司 一种车辆定位方法及系统
CN108682029A (zh) * 2018-03-22 2018-10-19 深圳飞马机器人科技有限公司 多尺度密集匹配方法和系统
CN109308737A (zh) * 2018-07-11 2019-02-05 重庆邮电大学 一种三阶段式点云配准方法的移动机器人v-slam方法
CN109282822B (zh) * 2018-08-31 2020-05-05 北京航空航天大学 构建导航地图的存储介质、方法和设备
CN109270545B (zh) * 2018-10-23 2020-08-11 百度在线网络技术(北京)有限公司 一种定位真值校验方法、装置、设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018126083A1 (en) * 2016-12-30 2018-07-05 DeepMap Inc. Alignment of data captured by autonomous vehicle to generate high definition maps
CN108345008A (zh) * 2017-01-23 2018-07-31 郑州宇通客车股份有限公司 一种目标物检测方法、点云数据提取方法及装置
CN108228798A (zh) * 2017-12-29 2018-06-29 百度在线网络技术(北京)有限公司 确定点云数据之间的匹配关系的方法和装置
CN109443369A (zh) * 2018-08-20 2019-03-08 北京主线科技有限公司 利用激光雷达和视觉传感器融合构建动静态栅格地图的方法
CN109814572A (zh) * 2019-02-20 2019-05-28 广州市山丘智能科技有限公司 移动机器人定位建图方法、装置、移动机器人和存储介质

Also Published As

Publication number Publication date
CN112154303A (zh) 2020-12-29
CN112154303B (zh) 2024-04-05

Similar Documents

Publication Publication Date Title
WO2022156175A1 (zh) 融合图像和点云信息的检测方法、系统、设备及存储介质
CN108921947B (zh) 生成电子地图的方法、装置、设备、存储介质以及采集实体
US11222204B2 (en) Creation of a 3D city model from oblique imaging and lidar data
EP3620823B1 (en) Method and device for detecting precision of internal parameter of laser radar
CN109270545B (zh) 一种定位真值校验方法、装置、设备及存储介质
CN108279670B (zh) 用于调整点云数据采集轨迹的方法、设备以及计算机可读介质
KR20190082070A (ko) 지도 생성 및 운동 객체 위치 결정 방법 및 장치
CN110703268B (zh) 一种自主定位导航的航线规划方法和装置
WO2021016803A1 (zh) 高精度地图定位方法、系统、平台及计算机可读存储介质
WO2018133727A1 (zh) 一种正射影像图的生成方法及装置
CN108519102B (zh) 一种基于二次投影的双目视觉里程计算方法
WO2020228694A1 (zh) 一种相机姿态信息检测方法、装置以及相应的智能驾驶设备
CN107527382B (zh) 数据处理方法以及装置
EP3716210B1 (en) Three-dimensional point group data generation method, position estimation method, three-dimensional point group data generation device, and position estimation device
CN112967344B (zh) 相机外参标定的方法、设备、存储介质及程序产品
JP2007218705A (ja) 白線モデル計測システム、計測台車および白線モデル計測装置
WO2022179094A1 (zh) 车载激光雷达外参数联合标定方法、系统、介质及设备
CN113409459A (zh) 高精地图的生产方法、装置、设备和计算机存储介质
WO2021016806A1 (zh) 高精度地图定位方法、系统、平台及计算机可读存储介质
CN113177974A (zh) 一种点云配准方法、装置、电子设备及存储介质
KR102490521B1 (ko) 라이다 좌표계와 카메라 좌표계의 벡터 정합을 통한 자동 캘리브레이션 방법
WO2024093635A1 (zh) 相机位姿估计方法、装置及计算机可读存储介质
WO2021051361A1 (zh) 高精度地图定位方法、系统、平台及计算机可读存储介质
CN109003295B (zh) 一种无人机航空影像快速匹配方法
Sani et al. 3D reconstruction of building model using UAV point clouds

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19940047

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19940047

Country of ref document: EP

Kind code of ref document: A1