Introduction to Supervised Classification

Creating an accurate map of forest cover is a fundamental task in forestry. **Supervised classification** is a powerful machine learning technique where we "teach" a model to recognize different land cover types. We do this by providing it with training data—examples of what "forest" and "non-forest" areas look like in a satellite image. The algorithm, in this case, **Random Forest**, learns the unique spectral signatures of these classes. It then uses this knowledge to classify every pixel in the entire image, resulting in a detailed land cover map. This method is highly effective for creating accurate baseline data for forest monitoring and management.

Analysis Workflow

Map Visualization

The final map displays the output of the Random Forest classification. Green areas represent pixels classified as 'Forest', while yellow areas represent 'Non-Forest'.

GEE Code Snippet

This script trains a Random Forest classifier on user-defined points to map forest cover from Sentinel-2 data and assesses its accuracy.

// Get Ground truth data
var forest = ee.FeatureCollection('users/your-username/forest_points'); // Example path
var nonforest = ee.FeatureCollection('users/your-username/nonforest_points'); // Example path

// Load Sentinel-2 image collection.
var s2 = ee.ImageCollection("COPERNICUS/S2");

// Define an area of interest (AOI).
var aoi = ee.Geometry.Polygon(
    [[[99.038, 13.380], [99.038, 12.553], [100.106, 12.553], [100.106, 13.380]]], null, false);

// Filter and preprocess the image collection.
var image = s2
  .filterBounds(aoi)
  .filterDate("2024-01-01", "2024-02-01")
  .filter(ee.Filter.lt("CLOUDY_PIXEL_PERCENTAGE", 5))
  .median()
  .clip(aoi);

// Prepare training data.
var trainingPoints = forest.merge(nonforest);

// Sample training data from the image.
var sample = image.sampleRegions({
  collection: trainingPoints,
  properties: ['landcover'],
  scale: 30
});

// Split data into training (70%) and validation (30%).
var sampleWithRandom = sample.randomColumn('random', 42);
var training = sampleWithRandom.filter(ee.Filter.lt('random', 0.7));
var validation = sampleWithRandom.filter(ee.Filter.gte('random', 0.7));

// Train a classifier.
var classifier = ee.Classifier.smileRandomForest(10).train({
  features: training,
  classProperty: 'landcover',
  inputProperties: image.bandNames()
});

// Classify the image.
var classified = image.classify(classifier);

// Assess accuracy.
var validationClassified = validation.classify(classifier);
var errorMatrix = validationClassified.errorMatrix('landcover', 'classification');
print('Error Matrix:', errorMatrix);
print('Overall Accuracy:', errorMatrix.accuracy());

// Define visualization parameters and display results.
var visParams = {min: 0, max: 1, palette: ["yellow", "green"]};
Map.centerObject(aoi, 9);
Map.addLayer(classified, visParams, "Classified Image");
Map.addLayer(trainingPoints, {color: 'red'}, "Training Points");

Analysis & Results