Research on the intelligent classifier of remote sensing image based on machine learning

It is based on the researches into the data mining technology centering on the method of machine learning. The authors hereby present an intelligent classifier of remote sensing image by upgrading the demonstrating manner of the remote sensing image categorizing, dealing with uncertainty and imprecision, perfecting the synaptic weights of the neural networks and optimizing the designing process of the topolopy structure. Initially, through the combination of neural networks and fuzzy logic with the utilization of equivalency between the two, the authors establish an equivalent fuzzy neural network model representing "IF-THEN" relationship, which is equipped with soft classifying capability of handling uncertainty and imprecision. And from this fuzzy neural network, the intelligent classifier is resulted by the utilization of the inquiry capability of the genetic algorithms into the entire space and the perfection of the synaptic weights and the network's topolopy structure of the model, rather than improve the quality of the network by deleting the neurons or connections of the network as a way to decrease the complexity of it.