RiverSnap - estimation of river hydraulic parameters using machine learning/AI models

Deep Learning based Water Segmentation in Satellite Imagery

This project includes data preprocessing, visualization, model training, and evaluation using TensorFlow and Keras, building a deep learning model for predicting binary water mask using UNet and pretrained models such as ResNet34, ResNet50 and EfficientNetV2B0, and establishing a UI using Flask, HTML, CSS & JS to give instant predictions.