Guided execution of hybrid similarity-measures for registration of partially overlapped aerial imagery

Abstract

This work presents a two-phase image registration technique utilizing a hybrid feature-based and an area-based similarity measure of partially overlapped aerial imagery in presence of affine translation and rotation transformations. The resulting selectively guided execution of similarity measures provides a reduction in search space, reducing the computational cost of the proposed algorithm. This multi-stage approach enhances the capability to perform image registration of low resolution imagery where scenes may have many structures but lack well defined structures for conventional feature extraction or lack to have enough variations in the intensity values to diminish statistical dependencies. The inherent statistical attributes of area-based methods are exploited through the sequential use of complex correlation and mutual information on physics-based features.