A New Approach for Dense Depth Image Recovery
Xiangfeng Zeng, Hao Fu, Bin Dai, Jun Hong Tan, Liang Xiao · 2015
The dense depth image recovery technology has attracted increasing attention in the field of Unmanned Ground Vehicle (UGV). A depth map at the resolution of a camera image simplifies image segmentation, object tracking, classification and recognition. This paper presents a method to generate dense depth image using bilateral filter framework and then refine it with Markov Random Field (MRF). Experimental results on the monocular images and sparse laser point clouds data from the KITTI raw datasets show that the dense depth image can be acquired under the complex scene in rural environments.