抄録
This paper introduces a fully automatic method for registering satellite images to vector maps in urban area. The key idea is to automatically match the man-made objects extracted from satellite images against the corresponding objects in maps. The proposed method consists of two procedures: feature extraction for matching, and determination of mismatch. We use the normalized difference vegetation index as a spectral indicator to extract matching features. The index is well known to provide accurate differentiation of vegetation from man-made objects. In order to minimize registration mismatch, we apply the voting technique based on the generalized Hough transform. The experimental results show that our approach can register satellite images against maps accurately; the registration mismatch is just a few pixels. We also applied our method to the multitemporal satellite image registration. It also offers reliable image-to-image registration.
