DhruvRaghav/Satellite_Image_Segmentation

Multi-Class Semantic Segmentation on India's Satellite Images.This project addresses the broader issue of semantic segmentation of satellite images by aiming at classifying each pixel as belonging to a Building & Road or not. We developed a Convolutional Neural Network suitable for this task, inspired from the U-net [7]. We trained our model on a set of two-dimensional satellite images. The corresponding labels were binary masks, ie. two-dimensional matrices with ones for pixels where a building was present, zeros otherwise. Given a satellite image as input, our network was then able to output a corresponding predicted binary mask. Image segmentation involves detecting and classifying individual objects within the image. Additionally, segmentation differs from object detection in that it works at the pixel level to determine the contours of objects within an image.