期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2016
卷号：4
期号：4
页码：7822
DOI：10.15680/IJIRCCE.2016.0404292
出版社：S&S Publications
摘要：Image registration is the first step towards using the satellite images for any purpose. It is the fundamental task used to match two or more partially overlapping multi view, multi modal or multi temporal images and stitches these images into one image comprising the whole scene. Automatic satellite image registration is a challenging task of overlaying two images for geometric conformity aligning common features by establishing a transformation model using distinguishable feature points collected simultaneously in reference image and the sensed images in a completely unassisted manner.Accuracy of registered image depends on accurate feature detection and matching. This paper aims to present a comparative analysis of feature detection techniqueslike Scale invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF), Harris Corner Detector and Improved version of Harris Corner Detectorused for automatic satellite image registration
