McGill University
M.Sc. candidate
Supervisor: Jeffrey Cardille
Start: 2014-09-08
End: 2016-08-31

Project

Bayesian Updating of Landcover Classification in Google Earth Engine
I am attempting to update old classification maps by using more recent satellite images which are quickly classified using unsupervised machine learning methods then combined the two classifications together using Bayesian statistics. Majority of the maps we use today are 5 - 10 years old because it takes that amount of time to create a map. The problem is 5 - 10 years is a long time especially with changes to the environment due to human development, natural disasters, and climate change. I am therefore trying to update older classification using more recent satellite images to create a map that has the same classification scheme but with more recent images.