Investigation of image fusion procedures using optimal registration and SVD algorithms

Abstract

Two types of optimization procedures are employed to both register and fuse image data. The registration problem utilizes a singular value decomposition (SVD) method. The actual fusion of the constituent images employs a maximum likelihood (ML) decision rule methodology. This ML procedure is modified through the square of the errors (Chi square) but provides an optimal method to accept and reject candidate images for inclusion into the fused images. Numerical simulations show the applicability of the method.