Positioning method and apparatus

1. A positioning method, comprising:

clustering points in a first point cloud through multi-clustering to obtain a second point cloud, wherein the second point cloud has a highest probability among the first point cloud of belonging to a target point cloud based on an average distance between the first point cloud and a centroid of the second point cloud according to a moving coordinate system of a self-driving apparatus, wherein the target point cloud represents a surface feature of a target object, and wherein the first point cloud comprises the target point cloud and another point cloud that represents a surface feature of an interfering object;

clustering points in a third point cloud at least once to obtain the target point cloud when the second point cloud is not the target point cloud, wherein the third point cloud is not the second point cloud; and

determining a position of the target object based on the target point cloud.

2. The positioning method of claim 1 , wherein clustering the points in a third point cloud at least once to obtain the target point cloud comprises:

clustering the points in the third point cloud to obtain a fourth point cloud, wherein the fourth point cloud has the highest probability, among the point clouds, of belonging to the target point cloud; and

clustering points in a fifth point cloud at least once to obtain the target point cloud when the fourth point cloud is not the target point cloud, a quantity of clustering operations is less than a first threshold, and a quantity of points in a fifth point cloud is not less than a second threshold, wherein the fifth point cloud is other than the fourth point cloud in the third point cloud.

3. The positioning method of claim 1 , wherein before clustering the points in the third point cloud at least once to obtain the target point cloud, the method further comprises extracting points other than the second point cloud in the first point cloud to obtain the third point cloud.

4. The positioning method of claim 1 , wherein clustering the points in the first point cloud to obtain the second point cloud comprises:

determining a clustering parameter based on a distribution of the first point cloud; and

clustering the points in the first point cloud using the clustering parameter to obtain the second point cloud.

5. The positioning method of claim 4 , wherein the clustering parameter is at least one of a neighborhood or a minimum quantity of points.

6. The positioning method of claim 1 , wherein before clustering the points in the first point cloud through multi-clustering to obtain the target point cloud, the method further comprises extracting the first point cloud from an original point cloud, wherein each point in the original point cloud corresponds to one pixel in an original image, wherein first pixels in the original image corresponding to the points in the first point cloud are within a target boundary of the original image, wherein second pixels in the original image corresponding to points other than the first point cloud in the original point cloud are not within the target boundary, and wherein an image of the target object is within the target boundary.

7. The positioning method of claim 6 , wherein extracting the first point cloud from the original point cloud comprises:

determining the target boundary in the original image; and

extracting the first point cloud from the original point cloud based on the target boundary.

8. The positioning method of claim 1 , further comprising:

obtaining, based on the position of the target object, a target driving operation for performance by the self-driving apparatus; and

sending a control instruction to the self-driving apparatus that instructs the self-driving apparatus to perform the target driving operation.

9. A positioning apparatus, comprising:

a non-transitory storage medium configured to store program instructions; and

a processor coupled to the non-transitory storage medium and configured to execute the program instructions that cause the positioning apparatus to:

cluster points in a first point cloud through multi-clustering to obtain a second point cloud, wherein the second point cloud has a highest probability among the first point cloud of belonging to a target cloud based on an average distance between the first point cloud and a centroid of the second point cloud according to a moving coordinate system of a self-driving apparatus, wherein the target point cloud represents a surface feature of a target object, and wherein the first point cloud comprises the target point cloud and another point cloud that represents a surface feature of an interfering object;

cluster points in a third point cloud at least once to obtain the target point cloud when the second point cloud is not the target point cloud, wherein the third point cloud is not the second point cloud; and

determine a position of the target object based on the target point cloud.

10. The positioning apparatus of claim 9 , wherein the program instructions further cause the positioning apparatus to:

cluster the points in the third point cloud to obtain a fourth point cloud, wherein the fourth point cloud has the highest probability of belonging to the target point cloud and that is in the plurality of point clouds; and

cluster points in a fifth point cloud at least once to obtain the target point cloud when the fourth point cloud is not the target point cloud, a quantity of clustering operations is less than a first threshold, and a quantity of points in a fifth point cloud is not less than a second threshold, wherein the fifth point cloud is other than the fourth point cloud in the third point cloud.

11. The positioning apparatus of claim 9 , wherein the program instructions further cause the positioning apparatus to extract points other than the second point cloud in the first point cloud to obtain the third point cloud.

12. The positioning apparatus of claim 9 , wherein the program instructions further cause the positioning apparatus to:

determine a clustering parameter based on distribution of the first point cloud; and

cluster the points in the first point cloud using the clustering parameter, to obtain the second point cloud.

13. The positioning apparatus of claim 12 , wherein the clustering parameter is at least one of a neighborhood and a minimum quantity of points.

14. The positioning apparatus of claim 9 , wherein the program instructions further cause the positioning apparatus to extract the first point cloud from an original point cloud, wherein each point in the original point cloud corresponds to one pixel in an original image, wherein first pixels in the original image corresponding to the points in the first point cloud are within a target boundary of the original image, wherein second pixels in the original image corresponding to points other than the first point cloud in the original point cloud are not within the target boundary, and wherein an image of the target object is within the target boundary.

15. The positioning apparatus of claim 14 , wherein the program instructions further cause the positioning apparatus to:

determine the target boundary in the original image; and

extract the first point cloud from the original point cloud based on the target boundary.

16. The positioning apparatus of claim 9 , wherein the program instructions further cause the positioning apparatus to:

obtain, based on the position of the target object, a target driving operation for performance by the self-driving apparatus; and

send a control instruction to the self-driving apparatus that instructs the self-driving apparatus to perform the target driving operation.

17. A computer program product comprising instructions that are stored on a non-transitory computer readable medium and that when executed by a processor, causes an apparatus to:

cluster points in a first point cloud through multi-clustering to obtain a second point cloud, wherein the second point cloud has a highest probability among the first point cloud of belonging to a target point cloud based on an average distance between the first point cloud and a centroid of the second point cloud according to a moving coordinate system of a self-driving apparatus, wherein the target point cloud represents a surface feature of a target object, and wherein the first point cloud comprises the target point cloud and another point cloud that represents a surface feature of an interfering object;

cluster points in a third point cloud at least once to obtain the target point cloud when the second point cloud is not the target point cloud, wherein the third point cloud is not the second point cloud; and

determine a position of the target object based on the target point cloud.

18. The computer program product of claim 17 , wherein before clustering the points in a first point cloud through multi-clustering to obtain a target point cloud, the instructions further cause the apparatus to extract the first point cloud from an original point cloud, wherein each point in the original point cloud corresponds to one pixel in an original image, wherein first pixels in the original image corresponding to the points in the first point cloud are within a target boundary of the original image, wherein second pixels in the original image corresponding to points other than the first point cloud in the original point cloud are not within the target boundary, and wherein an image of the target object is within the target boundary.

19. The computer program product of claim 18 , wherein the instructions to extract the first point cloud from the original point cloud further causes the apparatus to:

determine the target boundary in the original image; and

extract the first point cloud from the original point cloud based on the target boundary.

20. The computer program product of claim 18 , wherein the instructions to cluster the points in the third point cloud at least once to obtain the target point cloud further cause the apparatus to:

cluster the points in the third point cloud to obtain a fourth point cloud, wherein the fourth point cloud has the highest probability, among the point clouds, of belonging to the target point cloud; and

cluster points in a fifth point cloud at least once to obtain the target point cloud when the fourth point cloud is not the target point cloud, a quantity of clustering operations is less than a first threshold, and a quantity of points in a fifth point cloud is not less than a second threshold, wherein the fifth point cloud is other than the fourth point cloud in the third point cloud.

21. The computer program product of claim 18 , wherein the instructions further cause the apparatus to extract points other than the second point cloud in the first point cloud to obtain the third point cloud.

22. The computer program product of claim 18 , wherein the instructions to cluster the points in the first point cloud to obtain the second point cloud further cause the apparatus to:

determine a clustering parameter based on a distribution of the first point cloud; and

cluster the points in the first point cloud using the clustering parameter to obtain the second point cloud.

23. The computer program product of claim 17 , wherein the instructions that when executed by the processor further causes the apparatus to:

obtain, based on the position of the target object, a target driving operation for performance by the self-driving apparatus; and

send a control instruction to the self-driving apparatus that instructs the self-driving apparatus to perform the target driving operation.