摘要
本文通过对Harris算子进行改进,自适应地提取角点特征,然后基于角点特征进行由粗到精的匹配,完成影像的配准;并利用MODIS影像、TM影像和HJ-1卫星影像3种不同分辨率的遥感影像进行配准实验,结果表明该自动配准方法能够达到亚像素级的配准精度,是一种高精度的影像配准方法。
An automatic method based on comer features for the registration of remote sensing image was proposed in this paper. In this registration scheme, Harris operator was developed to detect comer features adaptively. Then the registration was conducted from roughly to fine. Experiments were carried out on MODIS images, TM images and HJ-1 images whose resolution are different. The results demonstrated that the proposed method could make registration accuracy up to sub-pixel, which can be a registration method of high precision.
出处
《测绘科学》
CSCD
北大核心
2013年第3期160-162,133,共4页
Science of Surveying and Mapping
基金
国家自然科学基金项目(40901286)
于超分辨率重建技术的城市居住用地强度与热环境空间分异及关系研究
关键词
影像匹配
自动配准
HARRIS算子
遥感影像
image matching
automatic registration
Harris operator
remote sensing image
