Analysis of Long-term Changes for Land Use and Land Cover using Machine Learning: A case study

Abstract

Manipur is one of the states in the northeastern part of India that has experienced a tremendous change in forest cover and a rapid increase in urbanization. Over the past two decades, this region has relatively changed in land surface features mostly due to anthropogenic activities. Detection of the changes in the land use land cover features helps to ensure sustainable development of the region. To achieve this goal, the present study uses remotely sensed data from sentinel-2A and Landsat 7,8 platforms for the period from the year 2000 to 2022. Machine learning algorithms have been proven to be useful in mapping the various land cover features. Here the land uses land cover (LULC) features are classified into six categories namely dense forest, open forest, agriculture, built-up area, water body, and barren land using random forest method. The classification method yielded an overall accuracy(OA) of 94.5, 93.32, 93.58, and 94.61% and a kappa coefficient index of 0.912, 0.925, 0.914, and 0.938 for 2000, 2008, 2016, and 2022 respectively. The results of the study indicate that the forest cover over Manipur has decreased significantly over recent years.