REGION-BASED CLASSIFICATION OF ELECTRO-OPTICAL IMAGES WITH NEURAL NETWORKS AND FUZZY LOGIC
I.M. Petrosyuk, Yuri Zaichenko · International Journal of Computing · 2014
This paper reports on a novel approach to the optical information processing for the hyperspectral remote sensing systems by means of developed unification algorithm of the two mathematical tools: the fuzzy logic and the neural network. New neuro-fuzzy classification algorithm for hyperspectral remote sensed images has been proposed. It is able to replace complicated empirical formulae, which require the knowledge of dependences of many input parameters that rapidly change during of range time and difficult for crisp determination.