Optimization of Stochastic Image Georeferencing Algorithms Using Mutual Information

The possibility of optimization stochastic image referencing algorithms using mutual information as a target function of image referencing quality is investigated. Three types of mutual information Renyi, Tsallis and Shannon are investigated. Experimental studies have shown that optimized algorithms based on Renyi mutual information have the best results in terms of convergence speed and accuracy of deformation parameter estimates when referencing multispectral and multi-temporal images.