使用图像分割的遮挡恢复立体匹配算法

Handling occlusions method for stereo correspondence using graph cuts

针对立体匹配在稀疏纹理、重复纹理、深度不连续和遮挡区域存在的问题,提出了一种高效的立体匹配算法.该算法主要由像素匹配代价计算和视差图全局优化2个步骤组成.为了大幅减少当前算法在场景深度不连续处所产生的过渡平滑现象和在稀疏纹理处产生的错误匹配,采用基于图像采样噪声无关的自适应权重加窗匹配算法.为了求解遮挡区域和不连续性区域的像素视差,使用遮挡和平滑惩罚代价来约束整幅视差图,并采用基于图像分割的能量最小化方法求取最优解.实验结果表明,相比于局部和全局算法,该算法可以更快且准确地计算稀疏纹理、不连续性和遮挡区域的像素视差.

For the difficult points of stereo correspondence in the areas with sparse textures, patterns, discontinuities or occlusions, an efficient algorithm is proposed. It mainly consists of two steps： pixel matching cost computation and global optimization of the disparity map sequentially. The first step adopts a special pixel matching algorithm with adaptive weights, which is insensitive to image sampiing, so that both over-smoothing problems in discontinuities and disparity errors in sparse textural areas caused by current methods can be sharply reduced. The second step can explicitly integrate both occlusion and discontinuity costs into the energy functions to regularize the disparity map, and the optimum can be solved rapidly by graph-cut based energy minimization. The final experimental results have verified, compared with local and global methods separately, the proposal is faster and more accurate to estimate disparities of the areas with sparse textures, discontinuities or occlusions.