Comparative Analysis and Implication of Hyperion Hyperspectral and Landsat-8 Multispectral Dataset in Land Classification

Remote sensing via hyperspectral imaging delivers the crucial earth’s surface information in narrow spectral bands, which may not be possible with multispectral imaging. The classification algorithms play a vital role in highlighting or categorizing the essential features of the earth’s surface with respect to spectral information and generate thematic maps for further processing in different applications. Therefore, it is essential to explore the impact of well-defined or emerging classifiers on hyperspectral and multispectral datasets. In the present work, the performance of various classifiers, i.e., support vector machine (SVM), feedforward neural networks (FF-NN) and maximum likelihood classifier (MLC), has been evaluated using Earth Observation-1 (EO-1) Hyperion and Landsat-8 Operational Land Imager and Thermal Infrared Sensor over a part of the North Indian states. The experimental outcomes have confirmed that the FF-NN classifier achieved higher accuracy (91.20% with Hyperion; 82% with Landsat-8) as compared to other classification methods, i.e., SVM (87.60% with Hyperion and 80% with Landsat-8) and MLC (84.40% with Hyperion and 72.40% with Landsat-8). This study is important in terms of exploring the potential of hyperspectral imaging with different classification algorithms in various emerging applications.