Probabilistic depth map fusion of Kinect and stereo in real-time

Abstract

This paper proposes a probabilistic framework for real-time depth map fusion of Kinect and stereo. By modeling the depth imaging process as a random experiment, we turn the depth map fusion into a problem of probability density function (pdf) estimation, and the problem can be further decoupled into four parts: fusion space, influence term, visibility term and confidence term. Strategies for each part of the framework are presented to perform real-time fusion of Kinect and stereo. Experimental results demonstrate the effectiveness of the method.