This project uses remote sensing techniques to extract the surface of lakes from stellite Earth observations and study the evolution of their surface with time. The project is based on Earth observations from the Copernicus Sentinel-2 mission (https://sentinels.copernicus.eu/web/sentinel/copernicus/sentinel-2). However, with a little bit of work, it could be adapted to other Earth observations. The GitHub repository is here.
The project is based on Python 3.10. I recommend to use conda and the provided
environment.yml file:
$ conda env create -f environment.yml
The project uses Copernicus Sentinel-2 at 60m resolution that are accessed via the Sentinel Hub API of the Copernicus database.
You first need to configurate a new sentinelhub SHConfig object. You can use
the provided sentinelhub_config.py script which will create the configuration
file ~/.config/sentinelhub/config.toml.
You will need to replace your_sh_client_id with your client ID and
your_sh_client_secret with your client secret. To obtain these credentials,
you can follow the instructions in the documentation of the Sentinel Hub API
here.
To extract the monthly lake surface over time:
$ python extract_lake_area_in_time.py --sampling monthly
By default, the script covers the period between 1st October 2016 and the current day. To cover the full 2017 year:
$ python extract_lake_area_in_time.py --sampling monthly --beginning '2017-01-01' --end '2017-12-31'
To plot the evolution of the lake surface in time:
$ python plot_area_evolution.py
The scripts to extract and plot the lake surface use a dictionary of lakes
(for instance data/lakes_mexico_catalogue.json). To visualise the location
of the lakes, with the evolution of their surface, a script allows to
create an interactive map:
$ python interactive_map.py
The html file can be seen and explored here
I welcome all changes/ideas/suggestion, big or small, that aim at improving the projects. Feel free to fork the repository on GitHub and create pull requests. Please report any bugs that you find here.