This is an academic project conducted as part of my Master's Degree program. It is a practical exercise implementing classification methods using satellite data collected following the 2010 Haitian Earthquake. In the aftermath of the natural disaster, many displaced individuals created temporary shelters using blue tarps which were visible in satellite imagery, the examination of which which ultimately aided responders in locating those in need.
This project focuses on the automation of locating tarps within satellite imagery. The data consists of the RGB color values for 63,241 pixels from these photos, and a variety of algorithms are implemented with the goal of correctly identifying which pixels correspond to blue tarps using this data.
This project was written entirely in RMarkdown. To view the fullly knitted output of the project please download the .zip and view earthquake-rproject.html. The source .rmd file is also available in the repository.