Global optimization for bidirectional stereo matching with occlusion handling

This paper presents a new global energy optimization method with occlusion handling for stereo matching. We present a new cost function for modeling the problem of stereo matching. A new data term revised from normal NCC algorithm and an occlusion term are constructed in our proposed cost function. In addition, a bidirectional stereo matching strategy is introduced to refine the disparity calculating results. In this paper, we compared the proposed method with state-of-the-art stereo matching algorithm quantitatively and qualitatively. Our experiments are carried out on both real-scene and Middlebury datasets. Experimental results demonstrate that our method evidently outperforms the others. Our method achieves good acceptable results under complex imaging condition, especially for some complex real-scene images.