Point cloud denoising systems and methods

1. A method, comprising:

determining a respective local coordinate system for each point of a point cloud;

determining a respective first adaptive-shape neighborhood for each point of the point cloud based on each respective local coordinate system;

performing filtering associated with each respective first adaptive-shape neighborhood to obtain a respective second adaptive-shape neighborhood for each point of the point cloud;

determining local estimates for points inside each of the second adaptive-shape neighborhoods;

determining a vector normal to a plane associated with at least one of the local estimates; and

aggregating the local estimates for each point of the point cloud based on the vector to obtain a denoised point cloud.

2. The method of claim 1 , further comprising generating the point cloud based on one or more visible-light images captured by a visible-light imaging device and one or more infrared images captured by an infrared imaging device.

3. The method of claim 1 , further comprising determining an adaptive directional size based on a first direction in the local coordinate system of a first point of the point cloud and a second direction in the local coordinate system of a second point of the point cloud, wherein the filtering is based on the adaptive directional size.

4. The method of claim 1 , wherein the respective local coordinate system of a point of the point cloud is based on principal axes of a predetermined number of nearest neighbors of the point.

5. The method of claim 1 , wherein at least one of the first adaptive-shape neighborhoods is an anisotropic adaptive-shape neighborhood, and wherein at least one of the second adaptive-shape neighborhoods is an anisotropic adaptive-shape neighborhood.

6. The method of claim 1 , further comprising:

densifying the point cloud to obtain a densified point cloud; and

associating the local estimates with the densified point cloud,

wherein the aggregating is based on the local estimates and the associating.

7. The method of claim 6 , wherein the densifying is based on normal vectors associated with each point of the point cloud and/or local point density associated with each point of the point cloud.

8. The method of claim 1 , wherein the determining the respective first adaptive-shape neighborhood for each point is based on local polynomial approximation intersection of confidence intervals (LPA-ICI) approach.

9. The method of claim 1 , wherein the filtering comprises weighted order statistics (WOS) filtering.

10. The method of claim 1 , wherein the aggregating is based on local polynomial approximation.

11. The method of claim 1 , further comprising:

splitting the point cloud into a plurality of sub-point clouds; and

selectively densifying one or more of the plurality of sub-point clouds based on a density threshold value.

12. The method of claim 1 , further comprising splitting the point cloud into a plurality of sub-point clouds, wherein the performing filtering comprises, for each sub-point cloud of the plurality of sub-point clouds, performing filtering of the sub-point cloud based on a noise level associated with the sub-point cloud.

13. An imaging system, comprising:

one or more processors; and

a non-transitory machine readable medium comprising instructions stored therein, which when executed by the one or more processors, cause the one or more processors to perform operations comprising:

determining a respective local coordinate system for each point of a point cloud;

determining a respective first adaptive-shape neighborhood for each point of the point cloud based on each respective local coordinate system;

performing filtering associated with each respective first adaptive-shape neighborhood to obtain a respective second adaptive-shape neighborhood for each point of the point cloud;

determining local estimates for points inside each of the second adaptive-shape neighborhoods;

densifying the point cloud to obtain a densified point cloud;

associating the local estimates with the densified point cloud; and

aggregating, based on the local estimates and the associating, the local estimates for each point of the point cloud to obtain a denoised point cloud.

14. The imaging system of claim 13 , further comprising at least one of a visible-light imaging device or an infrared imaging device.

15. The imaging system of claim 14 , wherein the infrared imaging device is a thermal imaging device.

16. The imaging system of claim 13 , wherein the operations further comprise generating the point cloud based on one or more visible-light images captured by a visible-light imaging device and one or more infrared images captured by an infrared imaging device.

17. The imaging system of claim 13 , wherein the operations further comprise determining a vector normal to a plane associated with at least one of the local estimates, wherein the aggregating is further based on the vector.

18. The imaging system of claim 13 , wherein the operations further comprise determining an adaptive directional size based on a first direction in the local coordinate system of a first point of the point cloud and a second direction in the local coordinate system of a second point of the point cloud, wherein the filtering is based on the adaptive directional size.

19. The imaging system of claim 13 , wherein the respective local coordinate system of a point of the point cloud is based on principal axes of a predetermined number of nearest neighbors of the point.

20. The imaging system of claim 13 , wherein the densifying is based on normal vectors associated with each point of the point cloud and/or local point density associated with each point of the point cloud.