图像相关匹配算法的快速实现

Fast Algorithm for Image Correlation Matching

结合分级关键点互相关迭代法与搜索空间标记法,设计了一种有效提高图像相关匹配速度的算法。其中分级关键点互相关迭代算法快速实现了由重要到不重要像素点的逐级迭代互相关匹配计算,算法在保证精度的前提下可以随时结束并输出相关匹配值;而搜索空间标记法则能快速排除掉大量参数空间内不可能匹配的点。实验证明,这种算法能在正确配准图像的前提下大幅度提高匹配速度。

In view of two approximate continuities in mutual correlation image matching process： the image gray level distribution continuity and the mutual correlation coefficient value in the parameter space continuity, the iterative algorithm is proposed to calculate the mutual correlation from hierarchical key points and the search space mark principle. By hier- archical key points algorithm, mutual correlation coefficients of the matching images can be iterative calculated from the important points to unimportant points in the images hierarchically, and the correlation coefficient can be obtained at any time with satisfactory precision. The search space mark principle enable us remove massive spots in parameter space impossible to match. The experiments show that the new algorithm can greatly enhance matching speed and achieve accurate matching results.