A matching algorithm based on disparity maps fusion is proposed.Firstly,on the basis of normalized cross correlation(NCC),various disparity maps are computed using several different matching window sizes.Then,for each disparity of each disparity maps,the confidence level is evaluated by a new confidence measure,which combined left right consistency(LRC)with signal to noise ratio(SNR).Finally,a new proposed disparity maps fusion strategy is used for formation of weighted disparity map in terms of confidence level.This disparity maps fusion strategy considers not only the confidence level of the disparity itself but also its neighbors.The algorithm has been applied to a pair of TanDEM-X spotlight stereo images. The results demonstrate that the accuracy of DEM generated with the proposed algorithm is improved from 11.28 m to 8.41 m and the gross errors are effectively reduced.%提出了一种基于视差图融合的匹配方法.首先,基于归一化互相关系数(normalized cross correlation,NCC),利用多个不同尺寸的匹配窗口分别进行匹配,获取相应的视差图；然后,提出了一种左右一致性(left right consistency,LRC)和信噪比(signal to noise ratio,SNR)相结合的置信测度,用来评价视差图中每个视差的置信水平；在此基础上,提出了一种视差图融合策略,该策略对上述多个匹配窗口获取的视差图进行加权融合,融合时既考虑了视差本身的置信水平,也兼顾了其邻域视差的影响.采用 TanDEM-X 的聚束立体影像进行试验,结果表明,本文方法能有效减少 DEM粗差点,DEM高程精度由11．28 m 提高到8．41 m.
展开▼
