Method and system for calibrating lidar point data
A system for calibrating light detection and ranging (Lidar) point data can include a computing device having a processor, and a memory configured to store instructions executed by the processor. The instructions are implemented to receive a plurality of 3D points from a Lidar sensor, and calibrate the Lidar point data of the Lidar sensor according to an operation of rotating the plurality of 3D points at an arbitrary angle and an operation of generating planar ground models by randomly extracting the plurality of 3D points rotated by the arbitrary angle.
1. A system for calibrating light detection and ranging (Lidar) point data, comprising a computing device,
wherein the computing device includes:
a processor; and
a memory configured to store instructions executed by the processor,
wherein the instructions are implemented to receive a plurality of three-dimensional (3D) points from a Lidar sensor, and calibrate the Lidar point data of the Lidar sensor according to an operation of rotating the plurality of 3D points at an arbitrary angle and an operation of generating planar ground models by randomly extracting the plurality of 3D points rotated by the arbitrary angle.
2. The system of claim 1 , wherein the instructions are implemented to further perform an operation of generating a 2.5D grid map including a plurality of grid cells using the plurality of 3D points rotated by the arbitrary angle and an operation of extracting ground candidate grid cells from the plurality of grid cells using a height difference between a 3D point having a minimum height value and a 3D point having a maximum height value included in each of the plurality of grid cells between the operation of rotating and the operation of generating planar ground models.
3. The system of claim 2 , wherein the plurality of generated planar ground models having different inclinations are generated by randomly extracting 3D points included in the extracted ground candidate grid cells.
4. A method of calibrating light detection and ranging (Lidar) point data, comprising:
rotating, by a processor, a plurality of 3D points received from a Lidar sensor at a first arbitrary angle to randomly extract a plurality of first 3D points rotated by the first arbitrary angle and generate a first planar ground model;
rotating, by the processor, the plurality of 3D points at a second arbitrary angle to randomly extract a second a plurality of second 3D points rotated by the second arbitrary angle and generate a second planar ground model; and
calibrating, by the processor, Lidar point data of the Lidar sensor according to the generated first planar ground model and the generated second planar ground model.
5. The method of claim 4 , further comprising:
generating, by the processor, a first 2.5D grid map including a plurality of first grid cells using the plurality of first 3D points rotated by the first arbitrary angle; and
extracting, by the processor, first ground candidate grid cells from the plurality of first grid cells using a height difference between a 3D point having a minimum height value and a 3D point having a maximum height value included in each of the plurality of first grid cells,
wherein the first planar ground model is generated by randomly extracting the plurality of first 3D points included in any one of the first ground candidate grid cells.
6. The method of claim 4 , further comprising:
generating, by the processor, a second 2.5D grid map including a plurality of second grid cells using the plurality of second 3D points rotated by the second arbitrary angle; and
extracting, by the processor, second ground candidate grid cells from the plurality of second grid cells using a height difference between a 3D point having a minimum height value and a 3D point having a maximum height value included in each of the plurality of second grid cells,
wherein the second planar ground model is generated by randomly extracting the plurality of second 3D points included in any one of the second ground candidate grid cells.
7. The method of claim 4 , further comprising:
rotating, by the processor, a plurality of 3D points at a third arbitrary angle; and
generating, by the processor, a third planar ground model by randomly extracting a plurality of third 3D points rotated by the third arbitrary angle,
wherein a planar ground model having a small inclination is a planar ground model having the smallest inclination among the first planar ground model, the second planar ground model, and the third planar ground model.
8. The method of claim 4 , wherein the calibrating of, by the processor, the Lidar point data of the Lidar sensor according to a planar ground model having a small inclination includes:
storing, by the processor, a rotation angle corresponding to the planar ground model having the small inclination;
receiving, by the processor, a plurality of different 3D points from the Lidar sensor at different times;
rotating, by the processor, the plurality of different 3D points at the stored rotation angle;
generating, by the processor, other 2.5D grid maps including a plurality of different grid cells using a plurality of different 3D points rotated by the stored rotation angle;
extracting, by the processor, other ground candidate grid cells from the plurality of different grid cells;
randomly extracting, by the processor, the plurality of different 3D points included in any one of the other extracted ground candidate grid cells to generate other planar ground models; and
calculating, by the processor, an offset of the other planar ground models, and calibrating positions of the plurality of different 3D points according to the calculated offset.
9. A computing device comprising:
a processor; and
a memory configured to store instructions executed by the processor,
wherein the instructions are implemented to receive a plurality of 3D points from a light detection and ranging (Lidar) sensor, and calibrate the Lidar point data of the Lidar sensor according to an operation of rotating the plurality of 3D points at an arbitrary angle and an operation of generating planar ground models by randomly extracting the plurality of 3D points rotated by the arbitrary angle.
10. The computing device of claim 9 , wherein the instructions are implemented to further perform an operation of generating a 2.5D grid map including a plurality of grid cells using the plurality of 3D points rotated by the arbitrary angle and an operation of extracting ground candidate grid cells from the plurality of grid cells using a height difference between a 3D point having a minimum height value and a 3D point having a maximum height value included in each of the plurality of grid cells between the operation of rotating and the operation of generating planar ground models.
11. The computing device of claim 10 , wherein the plurality of generated planar ground models having different inclinations are generated by randomly extracting 3D points included in the extracted ground candidate grid cells.