A Remote Sensing, GIS Based Study on LULC Change Detection by Different Methods of Classifiers on Landsat Data

Abstract

AbstractThe present study was conducted to detect landuse/landcover (LULC) by using different methods of classifiers, i.e., (i) SAM (spectral angle mapper) algorithm, (ii) MLC (maximum likelihood classifier) and (iii) MDC (minimum distance classifier) algorithm. Landsat 5 and 8 were used to classify the decadal LULC for the past 31 years (from 1989 to 2020) in the Imphal West district of Manipur, which is a semi-urban region. This region has rapidly urbanized in the past few decades, consequently affecting the LULC pattern, which triggered the need to carry out this study by integrating GIS (Geographical Information System) tools with RS (Remote Sensing) data. The Decadal LULC change showed a significant difference in the areas of build-up (increased by 18.32%), forest (decreased by 4.19%) and agriculture (decreased by 10.17%). Thus, it concludes that there is a possibility that agriculture and forest areas will decrease due to development in the built-up areas as the population in this region has increased considerably. This study also demonstrates the potential of GIS and RS for identifying the LULC using Landsat data with the same resolution and for proper utilization of future land planning policies & natural resources.KeywordsLULCGISRSChange detectionSAMMLCMDC