Abstract:

Concerning the high computation complexity and low efficiency in traditional stereo matching method based on Belief Propagation (BP), a fast convergent stereo matching algorithm based on Sum of Absolute Difference (SAD) algorithm and BP algorithm was proposed. Firstly, the SAD matching method was used as a similarity decision criterion to determine the initial disparity map. When the energy function was constructed, the initial parallax map was used as a limit of function to get the optimization of disparity distribution by BP. And when calculating the confidence level of each pixel in the optimization process, the algorithm only utilized the information translated from the neighboring pixels in an adaptive support window, while ignoring the impact of the pixels beyond the window, then the nodes of BP were reduced and the convergence speed was improved. The experimental results show that the proposed algorithm can reduce matching computation time by 50%-60% and improve efficiency while maintaining the matching accuracy, and the proposed algorithm lays the foundation for the real-time applications.

Key words: stereo matching, initial disparity, energy function, belief propagation(BP), fast convergence

摘要：

针对传统置信传播（BP）立体匹配算法运算次数较多、效率低下的问题，提出了一种基于像素灰度绝对误差和（SAD）和BP的快速收敛立体匹配算法。首先使用SAD作为代价函数来计算初始视差值，并将可靠视差值作为约束项加入全局算法BP的能量函数中，进行全局的能量函数的优化；然后在优化过程中更新计算每个像素点的置信度时，考虑当前像素点自适应大小邻域内像素点对它的信息传递，而忽略距离较远的像素点的影响，从而减少了置信传播节点数并提高了置信度收敛的速度。实验结果表明，提出的算法在保持相近匹配精度的前提下，运行时间减少了50%~60%，提高了立体匹配效率，为实时应用打下了基础。

关键词: 立体匹配, 初始视差, 能量函数, 置信传播, 快速收敛