satellites-internship

Semantic segmenation of crop fields

Getting started

• Clone repo

git clone ...
cd satellites-internship

• Install dependencies

pip install -r requirements.txt

• Start training Unet3D(define in train.py path to dataset)

python train.py

• Build web app image

docker build . --tag pastis

• Start web app

docker run -p 5000:5000 pastis

NOTE: If you wanna use Sentinel Hub API for inference you need to create/have account and paste CREDENTIALS(instance_id,sh_client_id,sh_client_secret) in app.py file code:

config = SHConfig()
config.instance_id = "<PUT YOUR CREDENTIALS HERE>"
config.sh_client_id = "<PUT YOUR CREDENTIALS HERE>"
config.sh_client_secret = "<PUT YOUR CREDENTIALS HERE>"
config.save()