ABSTRACT This study assesses the accuracy of land use and land cover (LULC) classification in the Jargo River Watershed, Mirzapur District, Uttar Pradesh, India, for the year 2022. The study utilized unsupervised and supervised techniques on Sentinel-2A satellite imagery to analyze specific LULC categories, including water bodies, agricultural land, built-up areas, bare ground and pasture. After classifying land use and land cover types, around 110 random points were generated in ArcGIS 10.2 and then converted to KML for viewing in Google Earth Pro. Google Earth Pro was used to validate the accuracy of classified pixels. The results showed an overall accuracy of 84% and Kappa coefficient of 78% for unsupervised classification using the ISO Cluster Classification method. For supervised classification, the accuracy was notably higher, with an overall accuracy of 89.2% and Kappa coefficient of 92%, using the Maximum Likelihood Classification method. Both methods exhibited nearly perfect Kappa criteria; however, supervised classification proved more accurate for the Jargo River Watershed than unsupervised classification. The results indicate that both methods are effective for LULC assessment, with supervised classification providing higher accuracy and thus greater reliability for environmental monitoring and management.
Sahu et al. (Wed,) studied this question.