Google Earth Engine Tutorial: Estimate Building Counts

Steps to Estimate Building Counts in Google Earth Engine

1. Set Up Your Environment

Create a GEE account and access the Code Editor. Ensure you have the Google Earth Engine API installed in your local environment if needed.

2. Load and Preprocess Satellite Imagery

Select an appropriate satellite dataset. For example, use the landsat/LC08/C01/T1_SR dataset for Landsat 8 imagery. Apply preprocessing steps like cloud masking and normalization.

var dataset = ee.ImageCollection("LANDSAT/LC08/C01/T1_SR")  
    .filterDate('2020-01-01', '2020-12-31')  
    .filter(ee.Filter.lt('CLOUD_COVER', 10));  
    var image = dataset.select(['SR_B.']).first().clip(geometry);

3. Define Training Data for Classification

Collect training samples from OpenStreetMap (OSM) or manual digitization to identify buildings. Use these samples to train a classifier.

var trainingData = ee.FeatureCollection("projects/your-project/training-buildings");  
    var classifier = ee.Classifier.randomForest(10).train({  
      features: trainingData,  
      classProperty: 'class',  
      inputProperties: ['SR_B1', 'SR_B2', 'SR_B3']  
    });

4. Apply Classification to Satellite Imagery

Use the trained classifier to identify buildings in the selected imagery.

var classified = image.classify(classifier);  
    var buildings = classified.select('classification').rename('buildings');

5. Extract Building Vectors

Convert the classified image to a polygon layer to extract building footprints.

var buildingPolygons = buildings.reduceToVectors({  
      geometry: geometry,  
      reducer: ee.Reducer.count(),  
      scale: 30,  
      maxPixels: 1e10  
    });

6. Visualize and Export Results

Add the classified layer to the map and export the results as a GeoJSON or CSV file.

Map.addLayer(buildings, {palette: ['red']}, 'Buildings');  
    Export.table.toDrive({  
      collection: buildingPolygons,  
      description: 'buildings_export',  
      folder: 'GEE_Exports',  
      fileFormat: 'GeoJSON'  
    });