| CPC G06T 17/20 (2013.01) [G06T 7/13 (2017.01); G06T 7/40 (2013.01); G06T 7/60 (2013.01); G06V 10/60 (2022.01); G06T 2207/10024 (2013.01); G06T 2207/20192 (2013.01)] | 20 Claims |
|
1. A method, comprising:
receiving, from a capturing device, one or more frames of depth data and color data of an object;
converting the color data to grayscale data;
generating, from the depth data, a 3D mesh representation of a 3D space, wherein the 3D mesh representation comprises vertices which define triangular surfaces of the 3D mesh representation, and the mesh representation comprises adjacent planes of an object;
identifying triangular surfaces in the 3D mesh representation, wherein for each triangular surface, two vertices are in one of the adjacent planes and an outlier vertex is in another of the adjacent planes;
determining a mesh-based line which represents a common edge of the adjacent planes based on the outlier vertices;
determining a texture-based line which represents the common edge based on the grayscale data; and
determining a final line which represents the common edge based on one or both of the mesh-based line and the texture-based line;
wherein the texture-based line is obtained by determining, for each of the outlier vertices, after the outlier vertex is moved to the mesh-based line, a corresponding set of pixels of the grayscale data, and determining a peak brightness gradient in the set of pixels, and wherein peak brightness gradients which are below a specified magnitude do not contribute to the texture-based line.
|