The Application of Improved Fuzzy ARTMAP Neural Network in Remote Sensing Classification of Land-use

Land-use change is an important area of global change research, rapid and accurate access to land-use temporal and spatial variation information is a key technology to study land-use change. In this paper proposed a method which utilizes the improved model of fuzzy ARTMAP network - simplified fuzzy ARTMAP neural network for remote sensing land-use classification, and Take the TM remote sensing image of Yiwu as an example to experiment, we compared the classification results with the traditional BP neural network classification results. Tests showed that the improved ARTMAP neural network improved the Accuracy of misclassification; it also shows that the structure of the improved fuzzy ARTMAP network is simple and need less training time. The Simplified Fuzzy ARTMAP network is an effective model to deal with high dimensional remote sensing image classification.