Methods and systems for deterministic calculation of surface normal vectors for sparse point clouds

A method comprising:

• receiving a point cloud;
• projecting the point cloud to a sparse range map;
• generating, from the sparse range map, a vertical kernel range map using a vertical kernel and a horizontal kernel range map using a horizontal kernel;
• obtaining a guidance range map that is a smoothed approximation of the sparse range map;
• completing the sparse range map to obtain a dense range map by:
• for a given empty entry in the sparse range map, computing a vertical gradient and a horizontal gradient using the guidance range map for a corresponding entry in the guidance range map;
• and
• filling in the given empty entry using a corresponding entry in the vertical kernel range map or in the horizontal kernel range map, based on a comparison between the vertical gradient and the horizontal gradient;
• computing a surface normal vector using the dense range map;
• and
• outputting the point cloud associated with the surface normal vector.

A non-transitory computer readable medium having instructions encoded thereon, wherein the instructions are executable by a processing unit of a computing system to cause the computing system to:

• receive a point cloud;
• project the point cloud to a sparse range map;
• generate, from the sparse range map, a vertical kernel range map using a vertical kernel and a horizontal kernel range map using a horizontal kernel;
• obtain a guidance range map that is a smoothed approximation of the sparse range map;
• complete the sparse range map to obtain a dense range map by:
• for a given empty entry in the sparse range map, computing a vertical gradient and a horizontal gradient using the guidance range map for a corresponding entry in the guidance range map;
• and
• filling in the given empty entry using a corresponding entry in the vertical kernel range map or in the horizontal kernel range map, based on a comparison between the vertical gradient and the horizontal gradient;
• compute a surface normal vector using the dense range map;
• and
• output the point cloud associated with the surface normal vector.