I'm learning segmentation of satellite images with U-Net. The dataset is from the DeepGlobe 2018 Land Cover Classification challenge.
The code is based on https://github.com/TarunKumar1995-glitch/land_cover_classification_unet repository. The author created a medium article at https://baratam-tarunkumar.medium.com/land-cover-classification-with-u-net-aa618ea64a1b.
Dataset is downloaded from https://www.kaggle.com/datasets/geoap96/deepglobe2018-landcover-segmentation-traindataset. Move the data folder to the repository after unzipping the file. It contains training_data and test_data subfolders.