THEORETICAL ISSUES OF SEARCHING AERIAL PHOTOGRAPHS: A BIRD'S EYE VIEW

Abstract

We review some pattern matching algorithms and techniques motivated by the discrete theory of image processing. The problem inspiring this research is that of searching an aerial photograph for all appearances of some object. The issues we discuss are digitization, local errors, rotation and scaling. We review deterministic serial techniques that were used for multidimensional pattern matching and discuss their strengths and weaknesses. We introduce the new geometric model that has been recently found to be extremely suitable for the theory of discrete image processing. We introduce some of the new algorithms developed in the geometric model and show their capability of efficiently solving hitherto open problems.