Using metaheuristics for Optimizing Satellite Image Registration

This paper deals with the optimization techniques used for Satellite Images Registration (SIR). Our approach besides requiring less execution time improves the quality of the results. The proposed approach is based on a newly concept called “points of interest of an image”. Once extracted such points will serve as inputs for the similarity estimator whose type is the Sum Square Differences (SSD). This procedure is used to evaluate digital images registration. Implementations of Simulated Annealing and a Genetic Algorithm are used to solve the SIR problems. Numerical experiments are conducted to assess the performances of the SA and GA.