基于Zernike矩的多时相遥感影像匹配方法研究

基金项目: 国家自然科学基金资助项目(40801213)；国家科技支撑计划资助项目(2008BAK49B04)；中国地质大学(武汉)青年教师基金资助项目

详细信息

Multi-temporal Remote Sensing Image Matching Base on Zernike Moment Matching

¹Faculty of Information Engineering,China University of Geoscience(Wuhan;³88 Lumo Road,Wuhan 430074,China;²Wuhan Geotechnical Engineering and Surveying Institutem 209 Wansongyuan Road,Wuhan 430072,China

Funds: 国家自然科学基金资助项目(40801213)；国家科技支撑计划资助项目(2008BAK49B04)；中国地质大学(武汉)青年教师基金资助项目

摘要: 针对参与变化检测影像的特点,归纳多时相影像匹配方法应满足的条件,提出了一种基于Zernike矩的匹配策略。在进行多时相影像匹配时,该方法对灰度的线性变化不敏感,而且可以匹配存在较大旋转角度的影像。试验表明,该匹配策略具有较高的匹配正确率。
- 变化检测 /
- 影像匹配 /
- 配准 /
- Zernike矩特征
Abstract: In this paper,we analyse change detection image pair characteristics and conclude the condition which multi-temporal remote sensing image should satisfy.Consequently,we propose a match method originating from Zernike moment.The method is non-sensitive to gray variation and rotation angle.The experiment proves that the matching method owns higher matching correcti rate.
- change detection /
- image match /
- registration /
- Zernike moment feature