Comparative study of MODIS, LANDSAT-8, SENTINEL-2B, and LISS-4 images for Precision farming using NDVI approach

Abstract

This study aims to understand the potential use and application of satellite images in analyzing the vegetation of a given area. It utilizes images from 4 sensors/satellites, namely MODIS (Terra), LANDSAT 8, SENTINEL 2B, and LISS 4 (Resourcesat-2). The area chosen for analysis is ‘Bangalore North, a taluk in the Bangalore district of Karnataka, India, where about 23% of the total area is used for agriculture. The images obtained are analyzed for the extent of data that can be extracted, individual spatial variability, and their relative application in precision farming and vegetation analysis. The maps are generated using NDVI (Normalised difference in Vegetation Index) approach. They are then categorized into 14 classes, after which the maps are analyzed using a histogram and by extracting pixel count for each class and comparing the results among the sensors/satellites used. It was found that, of the four sensors/satellites, LISS-4 is best suitable for precision farming and vegetation analysis as the map obtained has a higher and clear resolution, along with better spatial variability. In the absence of LISS-4, Sentinel - 2B was a better choice, and MODIS was unsuitable for this purpose.