论文部分内容阅读
本文将近年来兴起的小波理论应用于立体视觉技术，提出了一种新的立体匹配算法。该算法利用Ｈａｒｄｙ小波基函数将原始象对分解成一系列具有良好时（空）领域定位性的带通分量，以多通道带通分量的相位信息为匹配基元，从而将视差估计转化为相位差估计问题，整个系统采用由粗到细引导的多通道匹配结构，将匹配约束溶于估计算法之中。实验结果表明，该算法结构简单，抗噪力强，能快速有效地产生密集的视差场。
In this paper, the rising of wavelet theory in recent years applied to stereo vision technology, put forward a new stereo matching algorithm. The algorithm uses Hardy’s wavelet basis function to decompose the original image pair into a series of band-pass components with good time-space locality. The phase information of the multi-channel band-pass components is used as the matching primitive to transform the disparity estimation into phase difference To estimate the problem, the whole system adopts a multi-channel matching structure that is guided from coarse to fine, and the matching constraint is dissolved in the estimation algorithm. Experimental results show that the proposed algorithm has the advantages of simple structure, strong anti-noise ability and fast and effective generation of dense parallax fields.