Depth estimation based on data fusion of image sensor and depth sensor frames
A method of depth estimation, including, receiving an image frame, determining a relative depth map based on the image frame, receiving a sparse depth frame, preprocessing the sparse depth frame, determining a scale-adjusted relative depth map based on the relative depth map and the preprocessed sparse depth frame and fusing the relative depth map and the scale-adjusted relative depth map to produce an absolute depth map.
1 . A method of depth estimation, comprising:
receiving an image frame;
determining a relative depth map based on the image frame;
receiving a sparse depth frame from a sparse depth sensor;
preprocessing the sparse depth frame wherein the preprocessing includes at least one of noise reduction, hole reduction, median filtering, and occlusion reduction;
after preprocessing, extending a field of view of the sparse depth frame to the field of view of the image frame, reducing noise in the sparse depth frame, and enhancing resolution of the sparse depth frame, resulting in a sparse depth map;
determining a scale-adjusted relative depth map based on the relative depth map and the sparse depth map; and
fusing the relative depth map and the scale-adjusted relative depth map to produce an absolute depth map.
2 . The method of depth estimation of claim 1 , wherein the sparse depth frame is based on at least one of time-of-flight, light detection and ranging and structured light.
3 . The method of depth estimation of claim 1 , wherein the determining of the relative depth map is determined utilizing an encoder and decoder.
4 . The method of depth estimation of claim 1 , wherein the determining of the absolute depth map is determined utilizing an encoder and decoder.
5 . The method of depth estimation of claim 1 , wherein the fusing is performed by a fusion network.
6 . A method of depth estimation, comprising:
receiving an image frame;
determining a relative depth map based on the image frame;
receiving a sparse depth frame from a sparse depth sensor;
preprocessing the sparse depth frame wherein the preprocessing includes at least one of noise reduction, hole reduction, median filtering, and occlusion reduction;
after preprocessing, extending a field of view of the sparse depth frame to the field of view of the image frame, reducing noise in the sparse depth frame, and enhancing resolution of the sparse depth frame, resulting in a sparse depth map;
and
fusing the relative depth map and the sparse depth map to produce an absolute depth map.
7 . The method of depth estimation of claim 6 , further comprising extending a field of view of the sparse depth frame.
8 . The method of depth estimation of claim 6 , further comprising enhancing a resolution of the sparse depth frame.
9 . The method of depth estimation of claim 6 , wherein the sparse depth frame is based on at least one of time-of-flight, light detection and ranging and structured light.
10 . The method of depth estimation of claim 6 , wherein the preprocessing includes at least one of a noise reduction, a hole reduction, a median filtering and an occlusion reduction.