测绘学报

• 学术论文 • 上一篇    下一篇

基于Contourlet域Hausdorff距离和粒子群的多源遥感图像匹配

陈飒,吴一全   

  1. 南京航空航天大学
  • 收稿日期:2009-07-27 修回日期:2010-02-12 出版日期:2010-12-22 发布日期:2010-12-22
  • 通讯作者: 陈飒

Multi-source Remote Sensing Image Matching Based on Contourlet-domain Hausdorff Distance and Particle Swarm Optimization

  • Received:2009-07-27 Revised:2010-02-12 Online:2010-12-22 Published:2010-12-22

摘要: 为了进一步提高多源遥感图像的匹配精度和运算效率,本文提出了一种利用Contourlet变换、Hausdorff距离和改进粒子群的遥感图像匹配算法。在分别对目标图像和参考图像进行Contourlet分解的基础上,采用小波模极大值法提取低频图像的边缘信息,以LTS-HD作为图像匹配的相似性度量准则,并利用一种带极值扰动的简化粒子群优化算法对低频边缘图像进行匹配操作,得到粗匹配点;然后根据粗匹配点的位置反演计算到原始图像,进行精匹配,最终实现全分辨率情况下遥感图像的匹配。实验结果表明,该算法与目前常用的遥感图像匹配算法相比,不仅具有更高的匹配精度和运算效率,还有较强的鲁棒性。

Abstract: To further improve the accuracy and efficiency of multi-source remote sensing image matching, an algorithm based on contourlet transform, Hausdorff distance and improved particle swarm optimization was proposed in this paper. Firstly, the target image and reference image were decomposed to the low resolution image using contourlet transform. Then, wavelet modulus maxima algorithm was employed to extract the edges of the low-frequency subbands, and Least-trimmed-squares Hausdorff Distance(LTS-HD) was used as similarity measure for image matching. Meanwhile, the extremum disturbed and simple particle swarm optimization was introduced to get the rough matching results. The position of rough matching results were corresponded to the original image and then the matching between the higher resolution images could be implemented stepwise up to the full resolution images. The experimental results show that, compared with those of other existing sensing image matching methods, the proposed algorithm has the high accuracy, efficiency and strong robustness.