THE ELECTRO-OPTICS RESEARCH INSTITUTE
Recent Publications Online



Scene reconstruction from partially overlapping images with use of composite

William J. Chimitt, Jr.,* and Laurence G. Hassebrook
The ElectroOptics Research Institute
Department of Electrical Engineering, University of Kentucky, Lexington, Kentucky 40506-0046

Abstract

            In many image processing applications, it is desirable to combine partially overlapping images into one larger scene. We introduce a supervised correlation filter method that automatically detects overlapping image pairs and estimates proper alignment. We use on-line composite filter design to achieve rotation- and overlap-invariant image pairing. We present stochastic system analysis that provides closed-form expressions that predict correlation plane response distribution, defines an overlap-invariant filter bank architecture, and ob-tains the optimum training set size. The resulting high-level image processing method is numerically effi-cient, suited for optoelectronic implementation, and competitive with manual reconstruction in speed and ac-curacy.

William J. Chimitt, Jr., and Laurence G. Hassebrook, "Scene reconstruction from partially overlapping images with use of composite ," J. Opt. Soc. Am. A 16(9), 2124-2135 (Sep 1999)

PDF file entire paper