Land cover clustering and classification of satellite images

Abstract

Land cover classification refers to the process of using remote sensing data to categorize different types of land cover like vegetation, water bodies and soil. This is helpful for gaining key information about the surface of the Earth and for the future interactions between human activities and the environment. These predicted interactions lead to the development of sustainable land use practices along with the protection of natural resources. This paper deals with classifying the land cover using unsupervised and supervised methods. The unsupervised method includes land cover detection using a K-means clustering algorithm and the supervised classification is done using random forest classifier. The evaluation parameter values are calculated and compared for the input and output images.