URGENT- 4 Year Wembanad LULC training data set creation

I have a geo-referenced shapefile of the Wembanad wetland in Kerala together with the satellite imagery for four separate years. What I now need is a set of supervised Land-Use / Land-Cover maps—one for each year—built inside a standard GIS environment. The classification must at minimum separate Urban areas, Water bodies and Vegetation, plus the additional classes I already identified (barren land, crops) and any other logical subclasses we agree on after a quick review of the imagery. You are free to suggest refinements as long as the core categories remain intact and the final legend stays consistent across all four years. All processing should be done in mainstream GIS software (ArcGIS, QGIS or an equivalent that can export common formats). Please keep a clean, fully documented workflow so that the steps can be reproduced later if the study is expanded. Deliverables • Four georeferenced LULC rasters (one per year) in GeoTIFF • A concise methodology report summarising training-sample selection, classification algorithm, accuracy assessment and any post-processing you apply I will supply the shapefile boundaries and the satellite scenes as soon as we start; everything else—from training data(atleast 50 samples for each class) creation to accuracy checking—will be up to you. If the output aligns with the agreed legend and reaches acceptable overall accuracy, the job is complete.

Project ID: 40425808