Workshop: RS data processing with Python and Microsoft's Planetary Computer

This notebook was initially made for the "Remote Sensing" module (EPF MDE, Montpellier, FR).
Open an issue if you have any code related question (with reproductible the code snippet and error traceback).

Startup

Clone this repository using a shell in the Planetary Computer Jupyterlab interface.
You can connect your PC hub instance to VSCode see the docs here.

If you don't have a Planetary Computer account yet
Ask it. You can also create a local conda env using the provided YAML file.

conda env create -f conda_env.yml

It should work without an account for some images, but I fear some data (16bit images) requires an API key.

Evaluation

Before the second session, find another area and change to analyze with index (dNBR, NDVI, NDWI, NDSI).
Explore other datasets like Landsat that allow you to go back earlier in time (see the catalog).
Create a notebook with commented functions for your pipeline + plots and maps to show off your results.

Some ideas:

• wildfires
• deforestation
• coastal erosion
• floods
• snow cover
• ...

Ressources

Planetary Computer

Docs
Catalog

More docs

GDAL
Matplotlib
Numpy
Rasterio
Leafmap
Xarray
Xarray-Spatial
StackSTAC

DB of remote sensing index

Index DataBase

Communities

Pangeo
Group on Earth Observations

Useful GitHub repositories

See topics https://github.com/topics/earth-observation and https://github.com/topics/satellite-imagery
GDAL cheatsheet
Rasterstats
SentinelSat
EODAG
PyTorch + Geo

Exhaustive lists of tools and datasets

https://github.com/sacridini/Awesome-Geospatial
https://github.com/chrieke/awesome-geospatial-companies
https://github.com/robmarkcole/satellite-image-deep-learning
https://github.com/NRCan/geo-deep-learning
https://github.com/Seyed-Ali-Ahmadi/Awesome_Satellite_Benchmark_Datasets
https://github.com/DahnJ/Awesome-DEM