摘要
针对有云遥感影像间常规匹配效果不佳的问题,提出了一种考虑云检测的遥感影像匹配方法,主要包括有云区域的探测、特征点提取和相位匹配三个部分。通过对影像局部区域的高亮度像素的统计,在特征点提取时避开有云区域,最后通过物方空间的相位相关获取同名点对。实验表明本方法在包含大量云团的影像之间能够明显改善匹配效果。
To overcome the matching difficulties between remote sensing images with cloud, we proposed a remote sensing image matching method considering cloud detection. This method mainly includes cloud area detection, feature point extraction and phase correlation. Through the statistics of high-brightness pixelsin the local area, we exclude the cloud regions during feature point extraction. We then obtain the corresponding point pairs by the phase correlation in the object space. The result shows that this method can obviously improve the matching effect between images with a large number of clouds.
出处
《地理空间信息》
2019年第4期15-17,I0002,共4页
Geospatial Information
基金
湖北省国土资源科研计划项目(鄂土资文﹝2018﹞844号)
关键词
云检测
特征点检测
影像匹配
相位相关
cloud detection
feature points detection
image matching
phase correlation
