基于互相关的多源遥感图像匹配的改进算法

基于互相关的图像匹配算法是图像匹配中最经典的算法之一.本文主要从理论上分析了互相关用于多源图像匹配中可能导致出现误匹配的若干问题,提出了基于负指数的图像变换方法,增强提供精确匹配信息的像素的权重,并针对互相关曲面的特性,定义了新的匹配度量,来选取准确的图像匹配位置.利用星栽光学和星载SAR图像对算法进行了验证,实验表明,改进的算法用于多源图像的粗匹配的性能较原算法有大幅的提升,使得相关匹配算法具有更广的应用范围.

Improved Algorithm for Multi-source Remote Sensing Images Based on Cross Correlation

The cross correlation based image registration algorithm is the most typical one of the existing image registration algorithms.Several problems of the cross correlation algorithm applying to multi-source image registration,which may cause mis-registration,are theoretically analyzed in this paper.Then the negative exponential transformation is applied to the image pair to enlarge the proportion of those pixels which provide information for accurate registration.Considering the characteristic of the cross cor...