Welcome to deforest’s documentation!

The DEnse FOREst Time Series (DEFOREST) tool is a method for detecting changes in forest cover in a time series of Sentinel-2 data.

The processing chain does the following:

• Takes Sentinel-2 data (L1C/L2A) as input.
• Trains a model to robustly identify forest/nonforest across any phenological state.
• Classifies each image to a probability of forest cover.
• Combines these images under a Bayesian framework to identify forest cover change.
• Outputs maps of forest cover change, and ‘early warnings’ of change in recent imagery.

Contents:

• 1. Setup instructions
  • 1.1. Requirements
  • 1.2. Installing Anaconda Python
  • 1.3. Setting up your Anaconda environment
  • 1.4. Installing sen2mosaic
  • 1.5. Installing deforest
  • 1.6. Where do I get help?
• 2. Using deforest on the command line
  • 2.1. Downloading and Preprocessing Data
  • 2.2. Calibrating SMFM deforest
  • 2.3. Classification and change detection
• 3. Worked example on the command line
  • 3.1. Preparation
  • 3.2. Data preparation
  • 3.3. Training the classifier
  • 3.4. Classifing the data
  • 3.5. Change detection