A Modified Neural Network for Land use Land Cover Mapping of Landsat-8 Oli Data

With the advancement of technologies, now it is possible to capture the images of Earth’s topology and analyzing them with various tools and methods for understanding the changes as well to predict the topology for the future. In this work, the aim is to map the Land Use Land Cover (LULC) of the existing terrestrial features over the city Prayagraj, India with the help of modified Neural Network (NN),with 6-256-5 architecture and 3 hidden layers.The modified NN method is efficiently classifying the LULC features viz.; water, sand, barren, vegetation and urban with better individual class accuracy where the overall accuracy turns out to be 95.85 %. Overall accuracy of the classification increased as compared to previous classification results.