Classification of multispectral imagery using dynamic learning neural network

This paper presents the results of classification of SPOT high resolution visible (RHV) multispectral imagery using neural network. The test site, located near Taoyuan county of the northern Taiwan, is an agriculture area containing small ponds, bare and barren soils, vegetation, built-up land, near shore sea, and man-made buildings. The classifier is a dynamic learning neural network (DL) using the Kalman filter technique as adaptation rule. The network architecture is the multi-layer perceptrons, i.e., feed-forward nets with one or more layers of nodes between the input and output nodes. Methodology of selection of training data sets is addressed. Then, accordingly, selected data sets from 512×512 pixels three-band image are used to train the neural nets to categorize different types of the land-cover. Both simulated and real images are used to test the classification performance. Results indicate that the DL substantially reduces the training time as compared to commonly used back-propagation (BP) trained neural network whose slow training process was shown to impede it from certain practical applications. As for the classification accuracy, the presented results also shown to be excellent. It is concluded that the use of dynamic learning network gives very promising classification results in terms of training time and classification accuracy. In particular, the proposed network significantly improves the practicality of the land-cover classification.