Comparative evaluation of indian remote sensing multi-spectral sensors data for crop classification

Abstract

Digital data from four multi spectral sensors having different spatial resolution and spectral channel, LISS-III, LISS-II, LISS-I and WiFS acquired from two Indian Remote Sensing satellite platform (IRS-1B, IRS-IC) were evaluated for wheat crop classification accuracy over central India. Classification accuracy was assessed at different spatial resolutions using common red and near infrared (nir) bands in the above four sensors data as well as simulated intermediate resolution data. It was observed that accuracy improved with better spatial resolution and there was a considerable increases in accuracy from 188 to 144 m resolution and from 72 m up to 23 m resolution with the modest increase in the intermediate range. Inclusion of the MIR band, and choice of proper acquisition date resulted in significantly higher crop classification accuracies. The results have been evaluated against the future sensor planned on IRS Platform. © 2002 Taylor & Francis Group, LLC.