基于目标检测的SAR图像匹配算法
doi: 10.3724/SP.J.1146.2007.00920
A SAR Image Matching Algorithm Based on Target Detection
-
摘要: 该文提出一种基于目标检测的SAR图像匹配算法。针对SAR图像的特点，该算法先检测SAR图像的强散射目标，接着计算各强散射目标的质心，对主、辅图像的质心点集合进行Delaunay三角剖分，以三角剖分后的对应相似三角形的质心、内心、外心为匹配点进行匹配，最后用ERS-I，ERS-II卫星的SAR图像进行实验，通过与相干系数法进行比较，实验结果证实了方法的有效性。
-
关键词:
- SAR图像 /
- 目标检测 /
- Delaunay三角剖分 /
- 图像匹配
Abstract: This paper presents an algorithm about SAR image matching based on target detection．According to the characteristics of SAR image, this method detects the strong scatterer targets firstly，and then calculates the centroid points of the strong scatterers targets, then these centroid points are used to generating Delaunay triangulation, After triangulation，in the centroids, incenters, circumcenters of the corresponding similar triangle for establishing the maping function between master-image and slave-image. Finally, the test carried out with SAR images of ERS-I, ERS-II satellite, compareing with the methods based on correlation, the experimental result confirms the effectiveness of this method. -
[1] Barbara Zitov, Jan Flusser. Image registration methods: asurvey[J].Image and Vision Computing.2003, 21(11):977-1000 [2] 甘荣兵. 合成孔径雷达对抗及目标检测技术研究. [博士论文],成都, 电子科技大学, 2005, 11. [3] 马小虎. 一种基于图的平面点集Delaunay三角剖分算法, 中国图象图形学报, 1997年, 2(1), 7-10.Ma Xiao-hu. A graph-based algorithm for generating theDelaunay triangulation of a planar point set. China journal ofimage graphics, 1997, 2 (1), 7-10. [4] 黄勇. SAR图像变化检测及相关技术研究: [博士论文]. 成都,电子科技大学, 2006. [5] 张晓玲. 干涉合成孔径雷达成像处理技术研究. [博士论文],成都, 电子科技大学, 1999, 12. -
计量
- 文章访问数: 3322
- HTML全文浏览量: 81
- PDF下载量: 1165
- 被引次数: 0
下载:
