Land Use and Land Cover Classification Using Deep Belief Network for LISS-III Multispectral Satellite Images

Abstract

Land Use and Land Cover (LULC) classification is one of the familiar applications of geographical monitoring. Deep learning techniques like deep belief networks (DBN), are used for the purpose of feature extraction and classification of multispectral images. In this proposed framework, by applying DBN, spatial and spectral features were extracted and classified with high level of classification accuracy. LISS III images of Kottayam district, Kerala were used as experimental images. This proposed framework proved that, DBN has a high ability to extract the feature and classify the multispectral images with high accuracy than traditional methods.