Land Use/ Land Cover Classification of Google Earth Imagery

Abstract

Google Earth is a source of high spatial resolution images. The freely available Google Earth (GE) images are utilized to generate Land use/Land cover thematic map of the highly heterogeneous landscape of typical urban scene. In this paper, we have presented Euclidean Distance and Average Pixel Intensity based K-NN classification to classify five different land objects. The classification accuracy of the proposed method is compared against generic K-NN. The overall classification accuracy and the kappa value of generic K-NN are found to be 75.04% and 0.74 respectively. Whereas, proposed method results with 76.38% and 0.78. Both the methods exhibits classification error because of poor spectral reflectance properties of google earth imagery.