Introduction

Welcome to Land Cover Mapping Using Multiple Sensors! This workshop will demonstrate a Random Forest classification mapping workflow using a merged Landsat archive. We will cover Image Collection filtering, merging, compositing, and finally, Random Forest model training, classification and accuracy assessment.

Pre-workshop Set-up

1. Ensure you are logged into your Google Earth Engine account.
2. Click this link to accept the trinidad-tobago GEE script repository - https://code.earthengine.google.com/?accept_repo=users/kwoodward/caribbean-trainings
3. Create a new script file in your own script repository - name it ‘Land Cover Classification - Landsat’. Keep in mind the master copy is available in the caribbean-trainings script repository.

Objectives

1. Understand the general process for training and applying a model in Google Earth Engine on satellite data
2. Adapt the provided workflow for a different area of interest and time period
3. Experiment with different ways to improve accuracy of your classification