SIH (Smart India Hackathon) problem statement given by ISRO ======> NM-371 : Automated Mapping of Farms/Trees in a given Satellite Image of 5m Resolution.
My Machine Learning Model uses Semantic Segmentation Techinques using FCNN (Fully Convolutional Neural Network) architecture to map farmland regions and tree regions in a given satellite image having IR, red and green channels.
Using this method, I'm able to correctly map every pixel present in the image to the given classes.