Fractal method for detecting and filling data gaps within LiDAR data

A method of using a computer processor to improve the quality of a set of a three dimensional (3D) point cloud data representing a physical surface by detecting and filling null spaces, comprising:

• analyzing with said computer processor the set of 3D point cloud data to identify within said set of 3D point cloud data the presence of a plurality of level 1 fractals, each level 1 fractal respectively defined by a first plurality of voxels, the first plurality of voxels containing data points arranged in a first three dimensional pattern of a plurality of three-dimensional patterns;
• analyzing with said computer processor the set of 3D point cloud data to identify within said set of 3D point cloud data the presence of a plurality of level 2 fractals, each level 2 fractal defined by a second plurality of voxels, said second plurality of voxels comprised of at least a portion of two or more of said level 1 fractals, wherein said two or more level 1 fractals which form each level 2 fractal are also arranged in accordance with said first three dimensional pattern, whereby the three dimensional arrangement of the two or more level 1 fractals with respect to each other is the same as the three dimensional arrangement of the first plurality of voxels forming each level 1 fractal;
• and
• using said computer processor to selectively fill a third plurality of voxels in said 3D point cloud data with a first predetermined limited number of data points to increase a number of instances where said level 2 fractals can be used for representing said 3D point cloud data, and wherein a fractal self-similarity cost function is minimized.

A non-transitory computer-usable storage medium having stored thereon sequences of instructions that, when executed by a machine, cause the machine to perform actions which improve the quality of a set of a three dimensional (3D) point cloud data representing a physical surface by detecting and filling null spaces, the actions comprising:

• analyzing the set of 3D point cloud data with said computer processor to identify within said set of 3D point cloud data the presence of a plurality of level 1 fractals, each level 1 fractal defined by a first plurality of voxels containing data points arranged in a first three dimensional pattern of a plurality of three-dimensional patterns;
• and
• using said computer processor to selectively fill a second plurality of voxels in said 3D point cloud data with a first predetermined limited number of data points to increase a number of instances where level 2 fractals can be used for representing said 3D point cloud data, each level 2 fractal defined by a third plurality of voxels, said third plurality of voxels comprised of at least a portion of two or more level 1 fractals;
• and
• wherein said two or more level 1 fractals which form each level 2 fractal are also arranged in accordance with said first three dimensional pattern, whereby the three-dimensional arrangement of the two or more level 1 fractals with respect to each other is the same as the three dimensional arrangement of the first plurality of voxels forming each level 1 fractal.

A non-transitory computer-usable storage medium having stored thereon sequences of instructions that, when executed by a machine, cause the machine to perform actions which improve the quality of a set of a three dimensional (3D) point cloud data representing a physical surface by detecting and filling null spaces, the actions comprising:

• analyzing with said computer processor the set of 3D point cloud data to identify within said set of 3D point cloud data the presence of a plurality of level 1 fractals, each level 1 fractal respectively defined by a first plurality of voxels containing data points arranged in a first three dimensional pattern of a plurality of three-dimensional patterns;
• analyzing with said computer processor the set of 3D point cloud data to identify within said set of 3D point cloud data the presence of a plurality of level 2 fractals, each level 2 fractal defined by a second plurality of voxels, the second plurality of voxels comprised of at least a portion of two or more of said level 1 fractals, wherein said two or more level 1 fractals which form each level 2 fractal are also arranged in accordance with said first three dimensional pattern, whereby the three dimensional arrangement of the two or more level 1 fractals with respect to each other is the same as the three dimensional arrangement of the first plurality of voxels forming each level 1 fractal;
• and
• using said computer processor to selectively fill a third plurality of voxels in said 3D point cloud data with a first predetermined limited number of data points to increase a number of instances where said level 2 fractals can be used for representing said 3D point cloud data, and wherein a fractal self-similarity cost function is minimized.