Adaptive least squares matching as a non-linear least squares optimization problem

Adaptive Least Squares Matching (ALSM) is a powerful technique for precisely locating objects in digital images. The method was introduced to the photogrammetric community by Gruen in 1985 and has since been developed further. The purpose of this paper is to study the basic ALSM formulation from a least squares optimization point of view. It turns out that it is possible to describe the basic algorithm as a variation of the Gauss-Newton method for solving weighted non-linear least squares optimization problems. This opens the possibility of applying optimization theory on the ALSM problem. The line-search algorithm for obtaining global convergence is especially described and illustrate