Hierarchical Matching Techniques for Automatic Image Mosaicing
Charles Lee Begg
Department of Computer Science
University of Canterbury
Abstract
This paper looks at image mosaicing using hierarchical matching techniques as a way to create a novel view of a scene quickly. Several parameters including error function, region match and constraints are evaluated to find the fastest and most accurate mosaics. Constraints are shown to be very important to intensity matching. Hue matching is more accurate than intensity matching, and hue matching with the intensity correction produces overall better results.
