I'm using google earth engine to do some simple habitat classification and now want to export the result as a KML or KMZ. When I run this code it is layer 2 that I want to export - the habitat classification map.

  //sat image
    var S1 = ee.ImageCollection("COPERNICUS/S1_GRD"),
    S2 = ee.ImageCollection("COPERNICUS/S2"),
    fc = ee.FeatureCollection("ft:1jvw7bQoChrt-3dk4bQC96TMjZnnyrw6DvaLpfEdk"),//polygon
    fc2 = ee.FeatureCollection("ft:1_uzNGOXOEdfnRmkIYIkVCiMHGx8gCr-1ywGyhjLF"),//points
    fc3 = ee.FeatureCollection("ft:1WXJEHBhLLe1ztmsjzt7WyMe1Go04Uuj96XEQ-C0-");//study area
    //Set a point of interest 
    var aoi = fc; 
    //Set bands of interest
    var bands = ['B2', 'B3', 'B4','B8', 'B11', 'B12'];

//Import and filter Sentinel 2 collection
var S2 = ee.ImageCollection('COPERNICUS/S2')
.filterDate('2017-06-20', '2017-07-20')
var image = S2.mosaic();
// Clip the image to the polygon geometry
var landcover_roi = image.clip(fc3);

Map.addLayer(landcover_roi, {'bands': 'B12,B8,B4','min': 0,'max': 4000},        'Sentinel 2A SWIR,NIT,R', 0);
var training = landcover_roi.sampleRegions(fc2, ["name"], 59);
var classifier = ee.Classifier.cart()
.train(training, "name", landcover_roi.bandNames());
var out = landcover_roi.classify(classifier);
Map.addLayer(out, {palette: [
'ff0000',// not forest
'387242',// forest
'e0e0d1',// volcano 
], min:1, max:3});

// Create a geometry representing an export region.
var geometry = geometry;

Map.addLayer(ee.Image().paint(fc, 1, 3));
var fc = fc2;

 //Create a band for each veg type
 var class1 = out.eq(1).select([0],['not forest']);
 var class2 = out.eq(2).select([0],['forest']);
 var class3 = out.eq(3).select([0],['volcano']);
 // Concatanate the bands into a single image
 var total = ee.Image.cat([class1,class2,class3]);

 // Export the FeatureCollection to a KML file.
  image: total,
  description: 'imageToDriveExample',
  scale: 30,
  region: geometry
  • can you share the feature collections? – Rodrigo E. Principe Jul 24 '17 at 11:01
  • Not sure that I made one so perhaps that is the step I need to do. Everything in the code is shared and all the fusion table should be open. – James Jul 24 '17 at 11:07
  • Not to me. Permission denied Those are the training features.. may be you should try to make your own code and ask here what you don't understand or doesn't work as expected. – Rodrigo E. Principe Jul 24 '17 at 11:12
  • Sorry just changed the settings on the training points. This is my own code and everything works up until I need to export the results as a KML. – James Jul 24 '17 at 11:16

You could do it directly in GEE:

var shape = out.reduceToVectors({
  geometry: geometry,
  crs: ee.Projection('EPSG:4326'),
  scale: 10,
  maxPixels: 1e13

  collection: shape, 
  description: "yourDescription", 
  folder: "yourFolder", 

But I do not recommend to do it directly like that because there are a lot of small areas and it'll take too long to process (may be it won't do it). You could process it a bit more in GEE, like filtering by count property or apply a focal_mode or something, or you could download the raster and do the process locally with a raster to vector algorithm.

  • Rodrigo, thank you so much for your help. I really like the idea of downloading a raster. guessing I should repost this as another question asking how to download as a raster? – James Jul 24 '17 at 13:34
  • don't think so, in the last part of your code: Export.image.toDrive() just change total to out, and may be the scale should be 10 (Sentinel spatial resolution) – Rodrigo E. Principe Jul 24 '17 at 13:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.