在三维数字散斑整像素相关搜索中, 先利用极线约束和投影校正原理将相关搜索范围从整个图像约束到投影校正后的同一水平极线上, 再利用视差约束缩小搜索范围, 该相关搜索过程非常耗时。在投影校正和视差约束的基础上, 提出了基于灰度约束的三维数字散斑整像素相关搜索方法, 该方法可进一步滤除视差范围内大部分的待匹配点。当有效点个数为85783、相关窗口大小为9 pixel×9 pixel时, 相关搜索时间由7.24 s缩短为2.15 s。实验结果表明, 所提方法能有效提高三维数字散斑整像素相关搜索效率。

In the integer pixel correction searching for three-dimensional digital speckle, the epipolar constraint and the projective rectification principle are used to constrain the correlation searching range from the whole image to the same horizontal line after projection rectification, and then the disparity constraint is used to reduce the searching scope. The correlation searching process is time-consuming. A novel integer pixel correlation searching method for three-dimensional digital speckle based on gray constraint is proposed, which is based on projective rectification and disparity constraint. The proposed method can remove most of the candidate matching points in parallax scope. When the number of valid points is 85783, the size of the correlation window is 9 pixel×9 pixel, and the correlation searching time is reduced from 7.24 s to 2.15 s. Experimental results show that the efficiency of integer pixel correction searching for three-dimensional digital speckle can be improved greatly.