Mapping Coastal Green Infrastructure by Integrating Remote Sensing and Machine Learning in Cuddalore, Southeast Coast of the Bay of Bengal
Abstract:
Mapping Coastal Green Infrastructure (CGI) and mangrove habitats along the Cuddalore coast in Southeast India, using advanced remote sensing technologies, analyzed changes in land use, land cover (LULC), and the distribu tion of Normalized Difference Vegetation Index (NDVI) through satellite imagery from Sentinel-2, the Multispectral Instrument (MSI), Landsat, the Thematic Mapper (TM), the Enhanced Thematic Mapper (ETM+), the Thermal Infrared Sensor (TIRS), and the Operational Land Imager (OLI) for the years 2000 to 2024. This investigation reveals the significant growth in plantation areas, mangrove swamps, and agricultural land, alongside an expan sion of sand beaches and rivers. For mangroves, NDVI values are increased from 0.050 to 0.438, NDVI values shows good vegetation health, and a positive correlation between NDVI and temperature also highlights the rising temperatures in settlement areas. By using Google Earth Engine (GEE) with high-resolution Landsat and Sentinel data, CGI features were mapped with over 70% accuracy through a machine learning algorithm utilizing random forest techniques, with 30% validation from field surveys, particularly a random forest classifier. This study con f irms the ongoing degradation of coastal vegetation and shoreline erosion, emphasizing the need for targeted con servation strategies and offering scalable methodologies for addressing similar issues in global coastal regions.
