InterdisciplinaryProject

Repository for Interdisciplinary Project in DS at TU Wien

This project aims to develop a U-Net model for segmenting water bodies and monitoring the water extent of lake Neusiedl and its surrounding Lacken, using Sentinel-1 images.

Project Structure

• data_func.py & data_prep.ipynb - data preparation of my dataset for the model. Here the radar images are merged together, separate Lacken masks are created, and dataset is divided into train, validation and test set
• model.py - contains Encoder, Decoder and UNet classes
• model_v2.py - contains an UNet model class which incorporates a pre-trained model
• utils.py - contains the necessary functions needed for training the model and for calculating the metrics
• train&evaluate_model***.ipynb- the notebooks 1 to 4 are used for training the different models and evaluating them on test set
• inference.ipynb - the results of all the models are shown in one notebook through plots