Earth Engine: Can I export a feature collection as a tif?

Question

I am using Random Forest in Earth Engine to classify land use from Sentinel 2 images. I want to eventually have a raster file of the classification with each pixel classified using the RF model. At the moment my classification is a feature collection, can I convert/export this as a tif file?

Here is the code and the link to the code:

ee_code

In the code below "classified" is the clssification feature collection.

Map.addLayer(guan_ca);

var vis = sen2.filter(ee.Filter.calendarRange(2017,2019,'year'))
    .filter(ee.Filter.calendarRange(01,01,'month'))
    .filterBounds(guan_ca)
    .map(function(image){
    return image.clip(guan_ca);
});

var vis = vis.mean();

var trueColour = {
         bands: ["B4", "B3", "B2"],
         min: 0,
         max: 3000
     };

     // Add image to map
Map.addLayer(vis, trueColour, "true-colour image");
// Define false-colour visualization parameters
var falseColour = {
    bands: ["B8", "B4", "B3"],
    min: 0,
    max: 3000
};

// // Define false-colour visualization parameters
var falseColour = {
    bands: ["B8", "B4", "B3"],
    min: 0,
    max: 3000
};

// Add image to map
Map.addLayer(vis, falseColour, "false-colour image");


// // Create the training data

var newfc = ca.merge(water).merge(ca).merge(baresoil).merge(crops).merge(urban).merge(clouds).merge(shadows).merge(forest).merge(lightcover);

print(newfc);

var bands = ['B2', 'B3', 'B4', 'B5', 'B6', 'B7', 'B8']; 
var training = vis.select(bands).sampleRegions({
    collection: newfc,
    properties: ['landcover'],
    scale: 30
});    

print(training);

// Train the classifier, in this case, I use the Random Forest algorithm
var classifier = ee.Classifier.randomForest(100)
   .train({
           features: training,
           classProperty: 'landcover',
           inputProperties: bands
});

// Run the classification
var classified = vis.select(bands).classify(classifier);
print(classified);

// Display the classification
Map.centerObject(newfc, 11);
Map.addLayer(vis, 
{bands: ['B2', 'B3', 'B4'], max: 0.3}, 
'Sentinel 2 image');
Map.addLayer(classified, {min: 0, max: 8, palette: ['126CC6', 'EBEB07', '613D05', '2CBE17', 'B9BFB8', 'FFFFFF', '000000', '076105', 'D47606']}, 
'classification');
Map.addLayer(newfc); 

Export.table.toDrive({
  collection: newfc,
  description: ('landuse_trainingData'),
  fileFormat: 'KML',
});

Export.table.toDrive({
  collection: classified,
  description: ('land_use_cassification'),
  fileFormat: 'KML',
});

// Error estimation
var trainAccuracy = classifier.confusionMatrix();
print('Resubstitution error matrix: ', trainAccuracy);
print('Training overall accuracy: ', trainAccuracy.accuracy());

var withRandom = training.randomColumn('random');

// We want to reserve some of the data for testing, to avoid overfitting the model.
var split = 0.7;  // Roughly 70% training, 30% testing.
var trainingPartition = withRandom.filter(ee.Filter.lt('random', split));
var testingPartition = withRandom.filter(ee.Filter.gte('random', split));

// Trained with 70% of our data.
var trainedClassifier = ee.Classifier.randomForest().train({
  features: trainingPartition,
  classProperty: 'landcover',
  inputProperties: bands
});

// Classify the test FeatureCollection.
var test = testingPartition.classify(trainedClassifier);

// Print the confusion matrix.
var confusionMatrix = test.errorMatrix('landcover', 'classification');
print('RF testing error matrix', confusionMatrix);
print('RF testing accuraccy', confusionMatrix.accuracy());
print('RF testing accuracy', confusionMatrix.array());

var matrix = ee.Feature(null, {matrix: confusionMatrix.array ()});

Answer 1

ee.FeatureCollection.style() and ee.Image.paint() both can be used to draw a feature collection as an image, which you can then export.