Optimization-based depth repair and depth recovery

For improving the quality of depth information perception, depth repair (only based on depth information) and depth recovery (combined with other relevant information) are widely used in depth information pre-processing. In this paper, a weighted least square (WLS) framework was proposed to unify the depth repair and depth recovery. By effectively utilizing available information, the proposed method achieves better results of depth repair and recovery. The efficiency and effectivity of the proposed method have been fully validated in real systems.