Method, apparatus, device, and storage medium for calibrating posture of moving obstacle

1. A method for calibrating a posture of a moving obstacle, comprising:

obtaining a three-dimensional (3D) map, the 3D map comprising first static obstacles;

selecting a frame of point cloud data associated with the 3D map as a target frame of data, the target frame of data comprising second static obstacles and one or more moving obstacles;

determining posture information of each of the one or more moving obstacles in a coordinate system of the 3D map;

registering the target frame of data with the 3D map;

determining posture offset information of the target frame of data in the coordinate system according to a registration result;

calibrating the posture information of each of the one or more moving obstacles according to the posture offset information; and

adding each of the one or more moving obstacles after the calibrating into the 3D map,

wherein, registering the target frame of data with the 3D map, comprises:

obtaining one of the second static obstacles as a reference obstacle;

pairing the reference obstacle with the first static obstacles to obtain a pairing result:

performing an initial registration for the target frame of data in the 3D map according to the pairing result;

splitting one of the second static obstacles after the initial registration using a cube with a preset size, to form feature points with a preset granularity; and

performing a first precise registration for the target frame of data in the 3D map according to the obtained feature points,

wherein, pairing the reference obstacle with the first static obstacles to obtain the pairing result, and performing the initial registration for the target frame of data in the 3D map according to the pairing result, comprises:

determining a searching region in the coordinate system according to the posture information of the reference obstacle in the coordinate system;

searching a first static obstacle matching with the reference obstacle as a calibration obstacle according to the searching region; and

performing the initial registration for the target frame of data in the 3D map according to the posture information of the calibration obstacle in the coordinate system and the posture information of the reference obstacle in the coordinate system.

2. The method of claim 1 , further comprising:

performing a second precise registration for the target frame of data in the 3D map after the first precise registration using a point-to-point registration technique.

3. The method of claim 1 , further comprising:

deleting the one or more moving obstacles from the target frame of data.

4. The method of claim 1 , further comprising:

searching a plane associated with each of the one or more moving obstacles as a pasting plane in the 3D map; and

pasting a bottom of each of the one or more moving obstacles in the 3D map with the pasting plane.

5. The method of claim 4 , wherein, searching the plane associated with each of the one or more moving obstacles as the pasting plane in the 3D map, comprises:

in response to at least two planes comprised in a set region associated with a same moving obstacle in the 3D map, obtaining a movement track of a road collection device associated with the 3D map; and

determining the pasting plane from the at least two planes according to the movement track of the road collection device.

6. A device, comprising:

one or more processors; and

a memory configured to store one or more programs,

wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to perform:

obtaining a three-dimensional (3D) map, the 3D map comprising first static obstacles;

selecting a frame of point cloud data associated with the 3D map as a target frame of data, the target frame of data comprising second static obstacles and one or more moving obstacles;

determining posture information of each of the one or more moving obstacles in a coordinate system of the 3D map;

registering the target frame of data with the 3D map;

determining posture offset information of the target frame of data in the coordinate system according to a registration result;

calibrating the posture information of each of the one or more moving obstacles according to the posture offset information; and

adding each of the one or more moving obstacles after the calibrating into the 3D map,

wherein, registering the target frame of data with the 3D map, comprises:

obtaining one of the second static obstacles as a reference obstacle;

pairing the reference obstacle with the first static obstacles to obtain a pairing result:

performing an initial registration for the target frame of data in the 3D map according to the pairing result;

splitting one of the second static obstacles after the initial registration using a cube with a preset size, to form feature points with a preset granularity; and

performing a first precise registration for the target frame of data in the 3D map according to the obtained feature points,

wherein, pairing the reference obstacle with the first static obstacles to obtain the pairing result, and performing the initial registration for the target frame of data in the 3D map according to the pairing result, comprises:

determining a searching region in the coordinate system according to the posture information of the reference obstacle in the coordinate system;

searching a first static obstacle matching with the reference obstacle as a calibration obstacle according to the searching region; and

performing the initial registration for the target frame of data in the 3D map according to the posture information of the calibration obstacle in the coordinate system and the posture information of the reference obstacle in the coordinate system.

7. The device of claim 6 , wherein, when the one or more programs are executed by the one or more processors, the one or more processors are caused to further perform:

performing a second precise registration for the target frame of data in the 3D map after the first precise registration using a point-to-point registration technique.

8. The device of claim 6 , wherein, when the one or more programs are executed by the one or more processors, the one or more processors are caused to further perform:

deleting the one or more moving obstacles from the target frame of data.

9. The device of claim 6 , wherein, when the one or more programs are executed by the one or more processors, the one or more processors are caused to further perform:

searching a plane associated with each of the one or more moving obstacles as a pasting plane in the 3D map; and

pasting a bottom of each of the one or more moving obstacles in the 3D map with the pasting plane.

10. The device of claim 9 , wherein, searching the plane associated with each of the one or more moving obstacles as the pasting plane in the 3D map, comprises:

in response to at least two planes comprised in a set region associated with a same moving obstacle in the 3D map, obtaining a movement track of a road collection device associated with the 3D map; and

determining the pasting plane from the at least two planes according to the movement track of the road collection device.

11. A non-transient computer readable storage medium, stored thereon with computer programs that, when executed by a processor, perform a method, the method comprising:

obtaining a three-dimensional (3D) map, the 3D map comprising first static obstacles;

selecting a frame of point cloud data associated with the 3D map as a target frame of data, the target frame of data comprising second static obstacles and one or more moving obstacles;

determining posture information of each of the one or more moving obstacles in a coordinate system of the 3D map;

registering the target frame of data with the 3D map;

determining posture offset information of the target frame of data in the coordinate system according to a registration result;

calibrating the posture information of each of the one or more moving obstacles according to the posture offset information; and

adding each of the one or more moving obstacles after the calibrating into the 3D map,

wherein, registering the target frame of data with the 3D map, comprises:

obtaining one of the second static obstacles as a reference obstacle;

pairing the reference obstacle with the first static obstacles to obtain a pairing result:

performing an initial registration for the target frame of data in the 3D map according to the pairing result;

splitting one of the second static obstacles after the initial registration using a cube with a preset size, to form feature points with a preset granularity; and

performing a first precise registration for the target frame of data in the 3D map according to the obtained feature points,

wherein, pairing the reference obstacle with the first static obstacles to obtain the pairing result, and performing the initial registration for the target frame of data in the 3D map according to the pairing result, comprises:

determining a searching region in the coordinate system according to the posture information of the reference obstacle in the coordinate system;

searching a first static obstacle matching with the reference obstacle as a calibration obstacle according to the searching region; and

performing the initial registration for the target frame of data in the 3D map according to the posture information of the calibration obstacle in the coordinate system and the posture information of the reference obstacle in the coordinate system.

12. The non-transient computer readable storage medium of claim 11 , wherein the method further comprises:

performing a second precise registration for the target frame of data in the 3D map after the first precise registration using a point-to-point registration technique.

13. The non-transient computer readable storage medium of claim 11 , wherein the method further comprises:

deleting the one or more moving obstacles from the target frame of data.

14. The non-transient computer readable storage medium of claim 11 , wherein the method further comprises:

searching a plane associated with each of the one or more moving obstacles as a pasting plane in the 3D map; and

pasting a bottom of each of the one or more moving obstacles in the 3D map with the pasting plane.

15. The non-transient computer readable storage medium of claim 14 , wherein, searching the plane associated with each of the one or more moving obstacles as the pasting plane in the 3D map, comprises:

in response to at least two planes comprised in a set region associated with a same moving obstacle in the 3D map, obtaining a movement track of a road collection device associated with the 3D map; and

determining the pasting plane from the at least two planes according to the movement track of the road collection device.