Project for downloading and monitoring NYC satellite data from Sentinel 2 on an NVIDIA Jetson Nano. Using a ResNet-18 that can distinguish between 10 different kinds of land cover, characterize the change between the NYC region over time.
To start the container:
To run the container (i.e. then head to jupyter notebook and code):
sudo docker run --runtime nvidia -it --rm --network host -v ~/Projects/nyc-sentinel:/home/noah/nyc-sentinel nsk367/nyc-sentinel
Once inside the container, the ML model can be trained via
To download Sentinel Images, please refer to ''Sentinel Data Download Walkthrough.ipynb''
To analyze the different regions using the trained model, please refer to ''Land Pseudo Segmentation.ipynb''