Disclaimer: This repository was part of a university project in 2012. Scripts and expected data formats are likely outdated. Use at your own risk.
This collection of scripts was developed as a proof of concept during a university course for processing large Landsat data sets with R and automatically detect lakes. The goal is to automatically create binary lake masks so one can evaluate lake growth and lake distribution with minimal investment of time and money.
There are two main processing chains that can be achieved with these scripts.
- (crop) - calculate TOAR - create binary NDWI mask with threshold - (detect clumps to determine minimum lake size)
- (crop) - calculate TOAR - create NDWI Layers - detect threshold from training site and mask image - (detect clumps to determine minimum lake size)
Ideally each Landsat scene is in its own subfolder and each band is represented as a single GeoTiff (exactly as Landsat data is provided by the USGS).
This was a course project during the course Remote Sensing with R at the University of Erlangen-Nuernberg.