Exporting a classified image in GEE without getting a BLACK raster image

Question

I have been trying to export this classified image. I want my GeoTIFF file to show this view...

But every time I export, my image turns out Black. after some digging i got to learn that I need to attach the image with some rgb visual parameters with the export command, but i'm not sure how to do that. I have done this classification following a YouTube video as I'm learning. I'm a beginner in GEE.

The code I used is:

//CART classification and Accuracy Assessment
var image = ee.ImageCollection('LANDSAT/LC08/C02/T1_L2')
.filterDate('2016-01-01', '2020-12-31')
.filterBounds(roi)
.sort('CLOUD_COVER')
.first()
.clip(roi)


//Visualize
var visparamsTrue = {bands: ['SR_B4', 'SR_B3', 'SR_B2'], min:0, max:65455, gamma:1.4}
Map.addLayer(image, visparamsTrue, 'landsat_2017-2019')
Map.centerObject(roi)

//Create training data
var label = 'Class';
var bands = ['SR_B1', 'SR_B2', 'SR_B3', 'SR_B4', 'SR_B5', 'SR_B6', 'SR_B7']
var input = image.select(bands);

//merge
var training = urban.merge(water).merge(forest).merge(cropland).merge(barren)
print(training)

//overlay the points to get the training
var trainImage = input.sampleRegions({
  collection: training,
  properties: [label],
  scale: 30
})
print(trainImage)

//now we define how much(%) data we use for training
var trainingData = trainImage.randomColumn();
var trainSet = trainingData.filter(ee.Filter.lessThan('random', 0.8))
var testSet = trainingData.filter(ee.Filter.greaterThan('random', 0.8))

//create classification model
var classifier = ee.Classifier.smileCart().train(trainSet, label, bands)

//classify the image using upper created model
var classified = input.classify(classifier)
print(classified.getInfo())


//nown its to visulalize the data. first define a pallete color
var landcoverPalette = [
  '#253494', //For water(0)
  '#eff3ff', //for urban(1)
  '#31a354', //for forest(2)
  '#bae4b3', //for cropland(3)
  '#ffffd4', //for barren(4)];
];

//Add to maplayer
Map.addLayer(classified, {palette: landcoverPalette, min:0, max:4}, 'Classification')

// accuracy assessment
//classify the testset created earlier and get a confusin matrix
var confusionMatrix = ee.ConfusionMatrix(testSet.classify(classifier)
.errorMatrix({
  actual:'Class',
  predicted:'classification'
}));



print('confusionMatrix:', confusionMatrix)
//generate overall accuracy from confusionmatric
print('OverallAccuracy:', confusionMatrix.accuracy())

//Export
Export.image.toDrive({
  image: image,
  description: 'Classified_visualize_visparams',
  folder: 'CP2',  
  region: roi, scale: 30,
  maxPixels: 1e13,
}); 

I have tried exporting with this too, which gives me only the TRUE COLOR COMPOSITE view of this area, not the CLASSIFIED image which I'm looking for.

  Export.image.toDrive({
    image: image.visualize(visparamsTrue),
    description: 'Classified_visualize_visparams',
    folder: 'CP2',  
    region: roi, 
    scale: 30,
    maxPixels: 1e13,
    }); 

Answer 1

Because you are using true color in export visualization. Instead, use a classification palette.

var viz = { min:0, max:4, palette: landcoverPalette}

Export.image.toDrive({
    image: image.visualize(viz),
    description: 'Classified_visualize_visparams',
    folder: 'CP2',  
    region: roi, 
    scale: 30,
    maxPixels: 1e13,
    });