一种改进Hausdorff距离和自适应遗传搜索的遥感目标匹配方法

A remote sensing object matching approach based on an improved Hausdorff distance and autoadaptive genetic algorithm

采用快速点提取算子SIFT算子提取特征点,减少图像数据量,针对传统Hausdorff距离对噪声、出格点较敏感的问题对其进行改进,并以改进后的鲁棒Hausdorff距离作为匹配测度,利用非遍历而又有效的遗传搜索策略进一步提高了匹配速度。对发生旋转变形和灰度变化的遥感影像进行模拟实验,实验结果证明了该算法的有效性和快速性。

This paper proposed a fast image matching approach that could solve the problem of image rotation and intensity change. Firstly, a fast robust feature detector called SIFT detector is used to detect feature points in both test image and object image. Then, an improved Hausdorff distance used as matching measure is presented. Further more, an autoadaptive genetic algorithm as searching strategy is proposed. At last, experiment to satellite image that has been rotated and changed the intensity manually show that the proposed method in this paper is valid and superiority.