Land Use Land Cover Change Analysis Using Satellite Images With Deep Learning Models

Accurate understanding and prediction of Land Use and Land Cover (LULC) is crucial for better environmental management and sustainable development as they directly influence natural resource management, biodiversity, and the overall ecological balance. This work integrates remote sensing data and deep learning techniques to analyze and predict changes of land covers and vegetation dynamics over time. Using Resourcesat-1 LISS-3 imagery as the primary dataset for Resnet which gives the features to LSTM for change prediction analysis from 2009-2019. For the vegetation prediction analysis,Landsat data has been used to predict NDVI (Normalized Difference Vegetation Index) values of the years 2013 and 2023. The NDVI prediction utilizes an LSTM (Long Short-Term Memory) model designed to capture historical vegetation trends and forecast potential future changes. These predictions are then visualized in a time-lapse video format to illustrate seasonal shifts and long term changes in crop grown across regions. By taking these approaches, it is very useful for urban planning and agriculture development. Keywords:Remote Sensing Data, Deep Learning, Residual Neural Network (Resnet), LULC,LSTM