针对传统互信息图像匹配方法中因灰度概率分布计算量大而导致匹配速度慢的问题,提出一种增量式灰度统计的图像匹配新方法。该方法在初始待匹配子图概率分布统计表的基础上,经过增量式计算避免了相邻图像块中灰度概率统计的重复性操作,大大减少了图像匹配过程中的概率统计的运算量。实验结果证明了该算法在保持互信息图像匹配优势的同时,提高了图像匹配的速度。

To solve the problem of low matching speed resulting from large computational cost of gray probability distribution in traditional mutual-information-based image matching method, a new method of fast image matching based on incremental grayscale statistics was proposed. In this method, the probability distribution of to-be-matched sub-image was computed at first; and then incremental calculation was adopted to avoid repetitive operation of gray probability distribution between adjacent sub-images; Consequently, the amount of computation in the process of image matching was reduced. The experimental results demonstrate that the present algorithm not only keeps the advantages of image matching with mutual information, but also increases the speed of image matching.