Working with Glacial Lake Mapping Python Package (GLakeMap)

1. Installing machine learning packages

• Download and install Anacnonda software in your machine; links for details: https://www.anaconda.com/ https://www.anaconda.com/products/individual
• Open the Command Prompt window and type conda list. The command should list a lists of packages availabe in the environemnt. If the command exits with an error, add Anacnonda to the PATH or use Anaconda Prompt instead.
• Once conda command works, create a virtual environment and install Python (recommend Python 3.6) and the following packages:

   conda install -c anaconda numpy
   conda install -c anaconda scipy
   conda install -c anaconda matplotlib
   conda install -c anaconda scikit-learn
   conda install -c anaconda pandas
  

2. Installing arcpy package

• Install Background Geoprocessing (64-bit) software: https://desktop.arcgis.com/en/arcmap/latest/analyze/executing-tools/64bit-background.htm. Python 2.7 64-bit should be installed with this tool and arcpy package should also work.

3. Installing snappy package

• Download and install SNAP software in your machine: http://step.esa.int/main/download/snap-download/
• Configure Python to use the SNAP-Python (snappy) interface: https://senbox.atlassian.net/wiki/spaces/SNAP/pages/50855941/Configure+Python+to+use+the+SNAP-Python+snappy+interface
  • Note: Strongly recommend to use SNAP 64-bit version and Python 2.7 64-bit installed in step 2 for the latter.

4. Downloading data and using glakemap package

• Download Sentinel-1, Sentinel-2, and DEM data and place them, respectively, inside the folder data_directory which is provide along with the package. Downloading DEM data is not necessary. Check Process Dem Data section of the config.py file.
• Download glakemap package and set glakemap-dev as the current working directory. You can also clone the repository if you have Git installed in your machine. If so, use git clone command.
• Open and edit the config.py file. Directions for How To are provided as comments.
• Recomend to use Visual Studio Code : https://code.visualstudio.com/, for execcuting codes inside config.py. It is easy/handy to change between Python versions in VS Code. Spyder IDE can be used as well to excute the machine learning codes.
• Changing Python versions apply only when you have acees to ArcGIS Desktop

Possible Future Addition/Replacement

• Replace arcpy with gdal package for supporting Python 3 and above for all tasks in an open access manner.
• Add different machine learning models.

Query/Support

• Links here were provided to support you with relevant resources that I personally found useful while setting up proper and working environemnt.
• If you have any trouble or question, be it setting up the environemnt correcly or running the code successfully, drop me an email: s.wangchuk@vu.nl

References

• If you use this program in your work, please kindly refer the following publications:
  • Wangchuk, S., & Bolch, T. (2020). Mapping of glacial lakes using Sentinel-1 and Sentinel-2 data and a random forest classifier: Strengths and challenges. Science of Remote Sensing, 2, 100008.
  • Wangchuk, S., Bolch, T., & Zawadzki, J. (2019). Towards automated mapping and monitoring of potentially dangerous glacial lakes in Bhutan Himalaya using Sentinel-1 Synthetic Aperture Radar data. International journal of remote sensing, 40(12), 4642-4667.