Google Earth Engine – Converting Raster to Vector in Google Earth Engine

google-earth-enginevectorization

By running the code, one obtains:

FeatureCollection (Error) Image.reduceToVectors: First band ('x') of image must be integral.

I don´t understand this error. My goal is to convert the pixel values into vector data, e.g. Points.

var coorsfeat = mean.reduceToVectors({ reducer : null, geometry : geometry, crs :mean.projection().getInfo().crs, crsTransform :mean.projection().getInfo().transform, geometryType : 'centroid', labelProperty : 'Precip' })

Link to GEE

Best Answer

The error regarding the Integral was fixable by converting the input Data from float to Int. I recommend a multiplication to evade Data precision loss.

Fix: var mean= ImCol.filterBounds(aoi).filterDate('2017-01-01','2017-12-31').mean() .multiply(10000).toInt()

Unfortunately, this retrieves 0 Features.

Changing the Following Arguments to scale did the Trick :

crs :mean.projection().getInfo().crs, crsTransform :mean.projection().getInfo().transform

Unfortunately, this doesn't obtain one Feature per Pixel. Gonna open a new Thread for this one, as it seems to be a different problem then initially described in this thread

Related Solutions

Google Earth Engine – Raster to Vector Conversion in Google Earth Engine

Try not specifying the reducer. I think it's looking for additional bands to reduce (i.e. get the mean for a homogeneous region and set it as a property of the output feature) and not finding any.

Google Earth Engine – Converting Each Rastercell to Individual Point-Vector

Loosely based on this answer, you can make a feature collection consisting of points at the center of each pixel.

For your purpose, you can do that as follows:

// get the projection of the first image
var projection = ImCol.filterBounds(geometry).filterDate('2017-01-01','2017-12-31').first().projection();

function spacedPoints(AOI, projection) {
  // make a coordinate image
  // get coordinates image
  var latlon = ee.Image.pixelLonLat().reproject(projection);
  // put each lon lat in a list 
  var coords = latlon.select(['longitude', 'latitude'])
                 .reduceRegion({reducer: ee.Reducer.toList(),
                                geometry: AOI,
                                scale: projection.nominalScale().toInt(),
                                maxPixels: 10e10
  });
  // zip the coordinates for representation. Example: zip([1, 3],[2, 4]) --> [[1, 2], [3,4]]
  var point_list = ee.List(coords.get('longitude')).zip(ee.List(coords.get('latitude')));
  // preset a random list
  var list = ee.List([0]);
  // map over the point list and add features as geometry
    var feats = ee.FeatureCollection(point_list.map(function(point){
      var ind = point_list.indexOf(point);
      var feat = ee.Feature(ee.Geometry.Point(point_list.get(ind)), {'ID': ind});
    return list.add(feat);
    }).flatten().removeAll([0]));

  return feats;
}

// apply the function
var output = spacedPoints(geometry, projection); print(output)
Map.addLayer(output, {color: 'green'})

link script

Best Answer

Related Solutions

Google Earth Engine – Raster to Vector Conversion in Google Earth Engine

Google Earth Engine – Converting Each Rastercell to Individual Point-Vector

Related Question