Automatic sensor image registration using intelligent optimization methods

Sensor image registration is the task of correctly mapping one sensor image into another before sensor fusion. The problem is approached as an optimization problem, with a fitness function to evaluate the mapping parameter sets and several intelligent optimization methods, including genetic algorithms, simulated annealing, and genetic simulated annealing (GASA), used to automatically solve the problem. Experiments indicate the effectiveness of the registration methods which are insensitive to noise. GASA provides the best performance because of its rapid convergence.