Method and apparatus for detecting obstacle

1. A method for detecting an obstacle, comprising:

acquiring first point cloud data collected by a first vehicle-mounted laser radar and second point cloud data collected by a second vehicle-mounted laser radar, wherein the first vehicle-mounted laser radar and the second vehicle-mounted laser radar are mounted on a given autonomous driving vehicle, a height of the first vehicle-mounted laser radar from a ground is greater than a height of the second vehicle-mounted laser radar from the ground, and a number of wiring harnesses of the first vehicle-mounted laser radar is greater than a number of wiring harnesses of the second vehicle-mounted laser radar;

performing ground estimation based on the first point cloud data;

filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data; and

performing obstacle detection based on the second point cloud data after the ground point is filtered out,

wherein the performing ground estimation based on the first point cloud data comprises:

dividing the first point cloud data into a plurality of preset spatial grids, performing downsampling on the first point cloud data in each grid, and fitting a ground in the grid to obtain a first ground fitting result;

for each data point in a grid, calculating a distance between the data point and the fitted ground, using the data point as a new ground point in response to the distance being less than a threshold, and updating the first ground fitting result of the grid based on the new ground point to obtain a second ground fitting result;

calculating an angle between grounds in adjacent grids, smoothing the angle in response to the angle being smaller than a preset angle, and updating the second ground fitting result to obtain a third ground fitting result; and

obtaining the ground estimation result of the first point cloud data based on the third ground fitting result.

2. The method according to claim 1 , wherein the filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data comprises:

calculating a distance between a data point in the second point cloud data and the ground estimated based on the first point cloud data, and ascertaining a data point, of which a distance from the ground estimated based on the first point cloud data is less than a preset distance threshold, in the second point cloud data as the ground point; and

filtering out the ground point in the second point cloud data.

3. The method according to claim 1 , wherein the performing obstacle detection based on the second point cloud data after the ground point is filtered out comprises:

fusing the first point cloud data and the second point cloud data after the ground point is filtered out, to perform the obstacle detection.

4. The method according to claim 1 , wherein the second laser radar is a single-line laser radar.

5. An apparatus for detecting an obstacle, comprising:

at least one processor; and

a memory storing instructions, wherein the instructions when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising:

acquiring first point cloud data collected by a first vehicle-mounted laser radar and second point cloud data collected by a second vehicle-mounted laser radar, wherein the first vehicle-mounted laser radar and the second vehicle-mounted laser radar are mounted on a given autonomous driving vehicle, a height of the first vehicle-mounted laser radar from a ground is greater than a height of the second vehicle-mounted laser radar from the ground, and a number of wiring harnesses of the first vehicle-mounted laser radar is greater than a number of wiring harnesses of the second vehicle-mounted laser radar;

performing ground estimation based on the first point cloud data;

filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data; and

performing obstacle detection based on the second point cloud data after the ground point is filtered out,

wherein the performing ground estimation based on the first point cloud data comprises:

dividing the first point cloud data into a plurality of preset spatial grids, performing downsampling on the first point cloud data in each grid, and fitting a ground in the grid to obtain a first ground fitting result;

for each data point in a grid, calculating a distance between the data point and the fitted ground, using the data point as a new ground point in response to the distance being less than a threshold, and updating the first ground fitting result of the grid based on the new ground point to obtain a second ground fitting result;

calculating an angle between grounds in adjacent grids, smoothing the angle in response to the angle being smaller than a preset angle, and updating the second ground fitting result to obtain a third ground fitting result; and

obtaining the ground estimation result of the first point cloud data based on the third ground fitting result.

6. The apparatus according to claim 5 , wherein the filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data comprises:

calculating a distance between a data point in the second point cloud data and the ground estimated based on the first point cloud data, and ascertaining a data point, of which a distance from the ground estimated based on the first point cloud data is less than a preset distance threshold, in the second point cloud data as the ground point; and

filtering out the ground point in the second point cloud data.

7. The apparatus according to claim 5 , wherein the performing obstacle detection based on the second point cloud data after the ground point is filtered out comprises:

fusing the first point cloud data and the second point cloud data after the ground point is filtered out, to perform the obstacle detection.

8. The apparatus according to claim 5 , wherein the second laser radar is a single-line laser radar.

9. A non-transitory computer readable medium, storing a computer program, wherein the program, when executed by a processor, causes the processor to perform operations, the operations comprising:

acquiring first point cloud data collected by a first vehicle-mounted laser radar and second point cloud data collected by a second vehicle-mounted laser radar, wherein the first vehicle-mounted laser radar and the second vehicle-mounted laser radar are mounted on a given autonomous driving vehicle, a height of the first vehicle-mounted laser radar from a ground is greater than a height of the second vehicle-mounted laser radar from the ground, and a number of wiring harnesses of the first vehicle-mounted laser radar is greater than a number of wiring harnesses of the second vehicle-mounted laser radar;

performing ground estimation based on the first point cloud data;

filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data; and

performing obstacle detection based on the second point cloud data after the ground point is filtered out,

wherein the performing ground estimation based on the first point cloud data comprises:

dividing the first point cloud data into a plurality of preset spatial grids, performing downsampling on the first point cloud data in each grid, and fitting a ground in the grid to obtain a first ground fitting result

for each data point in a grid, calculating a distance between the data point and the fitted ground, using the data point as a new ground point in response to the distance being less than a threshold, and updating the first ground fitting result of the grid based on the new ground point to obtain a second ground fitting result;

calculating an angle between grounds in adjacent grids, smoothing the angle in response to the angle being smaller than a preset angle, and updating the second ground fitting result to obtain a third ground fitting result; and

obtaining the ground estimation result of the first point cloud data based on the third ground fitting result.

10. The non-transitory computer readable medium according to claim 9 , wherein the filtering out a ground point in the second point cloud data according to the ground estimation result of the first point cloud data comprises:

calculating a distance between a data point in the second point cloud data and the ground estimated based on the first point cloud data, and ascertaining a data point, of which a distance from the ground estimated based on the first point cloud data is less than a preset distance threshold, in the second point cloud data as the ground point; and

filtering out the ground point in the second point cloud data.

11. The non-transitory computer readable medium according to claim 9 , wherein the performing obstacle detection based on the second point cloud data after the ground point is filtered out comprises:

fusing the first point cloud data and the second point cloud data after the ground point is filtered out, to perform the obstacle detection.

12. The non-transitory computer readable medium according to claim 9 , wherein the second laser radar is a single-line laser radar.