Deep learning with TorchGeo and PyTorch Lightning using custom geodataset
Автор: geografif
Загружено: 2025-03-14
Просмотров: 810
Описание:
A basic yet straightforward tutorial on training (and testing) a geospatial artificial intelligence (GeoAI) model for satellite image semantic segmentation using PyTorch Lightning. TorchGeo is used to handle the georeferenced dataset of images and labels.
2-channels images used are VV and VH polarization images from Sentinel-1 C-band SAR. The labels are from publicly available oil palm maps.
Medium article: / geospatial-deep-learning-using-torchgeo-an...
Oil palm data: • Open data: Global oil palm plantation map 🌴
TorchGeo paper: https://arxiv.org/abs/2111.08872
Github: https://github.com/microsoft/torchgeo
#GeoAI
Повторяем попытку...
Доступные форматы для скачивания:
Скачать видео
-
Информация по загрузке:
