Exporting Images

You can export images from Earth Engine in GeoTIFF or TFRecord format. See Configuration Parameters for more output options.

Example Setup

Start by defining the image data that will be exported:

// Load a landsat image and select three bands.
var landsat = ee.Image('LANDSAT/LC08/C02/T1_TOA/LC08_123032_20140515')
  .select(['B4', 'B3', 'B2']);

// Create a geometry representing an export region.
var geometry = ee.Geometry.Rectangle([116.2621, 39.8412, 116.4849, 40.01236]);

import ee
import geemap.core as geemap

# Load a landsat image and select three bands.
landsat = ee.Image('LANDSAT/LC08/C02/T1_TOA/LC08_123032_20140515').select(
    ['B4', 'B3', 'B2']
)

# Create a geometry representing an export region.
geometry = ee.Geometry.Rectangle([116.2621, 39.8412, 116.4849, 40.01236])

Next define the projection parameters that will be used in the following exports. We use the crs parameter to specify the coordinate system, and the crsTransform parameter to precisely specify the pixel grid. The crsTransform parameter is a list of parameters from an affine transformation matrix in row-major order [xScale, xShearing, xTranslation, yShearing, yScale, yTranslation]. An image's origin is defined by the xTranslation and yTranslation values, and the image's pixel size is defined by the xScale and yScale values. See examples of affine matrices.

// Retrieve the projection information from a band of the original image.
// Call getInfo() on the projection to request a client-side object containing
// the crs and transform information needed for the client-side Export function.
var projection = landsat.select('B2').projection().getInfo();

import ee
import geemap.core as geemap

# Retrieve the projection information from a band of the original image.
# Call getInfo() on the projection to request a client-side object containing
# the crs and transform information needed for the client-side Export function.
projection = landsat.select('B2').projection().getInfo()

Setting scale

As a shortcut, you can specify a scale parameter and Earth Engine will calculate a crsTransform parameter for you. However, simply setting the scale of an image does not specify the origin of the projection, and may result in an image that is shifted relative to another image with the same pixel size!

The reason for the potential shift is that the scale parameter is used to populate the xScale and yScale values of the crsTransform, but the xTranslation and yTranslation values are calculated so that if they are divided by the corresponding xScale and yScale values the remainder will be zero. These parameters specify a pixel grid where the projection's origin is at the corner of a pixel. This convention differs from the translation parameters used by some data providers, which use grids that are offset from the projection's origin. For example, Landsat images provided by USGS use translation parameters that are offset by a 1/2 pixel from the projection's origin (15m offset for the 30m bands) while Sentinel-2 images provided by ESA use translation parameters that are aligned with the projection's origin. If the crsTransform specified in an export do not match the crsTransform of the original image, the output pixels will be resampled (using nearest neighbor by default), which will make the resulting image be shifted relative to the original image.

To sum up, if you need to match the exported image's pixels to a specific image, make sure to use the crs and crsTransform parameters for full control of the grid.

to Drive

To export an image to your Drive account, use Export.image.toDrive(). For example, to export portions of a Landsat image, define a region to export, then call Export.image.toDrive():

// Export the image, specifying the CRS, transform, and region.
Export.image.toDrive({
  image: landsat,
  description: 'imageToDriveExample_transform',
  crs: projection.crs,
  crsTransform: projection.transform,
  region: geometry
});

import ee
import geemap.core as geemap

# Export the image, specifying the CRS, transform, and region.
task = ee.batch.Export.image.toDrive(
    image=landsat,
    description='imageToDriveExample_transform',
    crs=projection['crs'],
    crsTransform=projection['transform'],
    region=geometry,
)
task.start()

When this code is run, an export task will be created in the Code Editor Tasks tab. Click the Run button next to the task to start it. (Learn more about the Task Manager from the Code Editor section). The image will be created in your Drive account with the specified fileFormat.

to Cloud Storage

To export an image to a Google Cloud Storage bucket, use Export.image.toCloudStorage(). To export the Landsat image in the previous example to Cloud Storage instead of Drive, use:

// Export the image to Cloud Storage.
Export.image.toCloudStorage({
  image: landsat,
  description: 'imageToCloudExample',
  bucket: 'your-bucket-name',
  fileNamePrefix: 'exampleExport',
  crs: projection.crs,
  crsTransform: projection.transform,
  region: geometry
});

import ee
import geemap.core as geemap

# Export the image to Cloud Storage.
task = ee.batch.Export.image.toCloudStorage(
    image=landsat,
    description='imageToCloudExample',
    bucket='your-bucket-name',
    fileNamePrefix='exampleExport',
    crs=projection['crs'],
    crsTransform=projection['transform'],
    region=geometry,
)
task.start()

As with exports to Drive, start the export from the Tasks tab. The Cloud Storage bucket location can affect performance and storage costs, see the FAQ entry on location considerations for more information.

to Asset

To export an image to an asset in your Earth Engine assets folder, use Export.image.toAsset(). To manage your Earth Engine assets, or check how much of your storage quota is in use, use the Asset Manager. The following example illustrates exporting portions of a Landsat image using different pyramiding policies for the same band. The pyramiding policy indicates how Earth Engine computes lower-resolution versions of the asset. Learn more about how Earth Engine handles multiple resolutions in the scale doc.

// Get band 4 from the Landsat image, copy it.
var band4 = landsat.select('B4').rename('b4_mean')
  .addBands(landsat.select('B4').rename('b4_sample'))
  .addBands(landsat.select('B4').rename('b4_max'));

// Export the image to an Earth Engine asset.
Export.image.toAsset({
  image: band4,
  description: 'imageToAssetExample',
  assetId: 'exampleExport',
  crs: projection.crs,
  crsTransform: projection.transform,
  region: geometry,
  pyramidingPolicy: {
    'b4_mean': 'mean',
    'b4_sample': 'sample',
    'b4_max': 'max'
  }
});

import ee
import geemap.core as geemap

# Get band 4 from the Landsat image, copy it.
band_4 = (
    landsat.select('B4')
    .rename('b4_mean')
    .addBands(landsat.select('B4').rename('b4_sample'))
    .addBands(landsat.select('B4').rename('b4_max'))
)

# Export the image to an Earth Engine asset.
task = ee.batch.Export.image.toAsset(
    image=band_4,
    description='imageToAssetExample',
    assetId='projects/your-project/assets/exampleExport',
    crs=projection['crs'],
    crsTransform=projection['transform'],
    region=geometry,
    pyramidingPolicy={
        'b4_mean': 'mean',
        'b4_sample': 'sample',
        'b4_max': 'max',
    },
)
task.start()

You can provide a default pyramiding policy for every band that isn't explicitly specified by using the '.default' key. You may also pass in just the '.default' key. For example, to make all bands default to the 'sample' pyramiding policy, use {'.default': 'sample'}.

Configuration parameters

Observe that the dictionary of configuration parameters passed to Export.image includes scale (in meters) and the export region as an ee.Geometry. The exported image will cover the specified region with pixels at the specified scale. If not explicitly specified, the CRS of the output will be taken from the first band of the image to be exported.

You may also specify the dimensions, crs and/or crsTransform of the exported image. See the glossary for more information on crs and crsTransform. For example, to get a block of pixels precisely aligned to another data source, specify dimensions, crs and crsTransform. To get a block of pixels of predefined size (for example a 256x256 thumbnail image) that covers a region, specify dimensions and region.

You can specify image output format (if the destination is not toAsset()) with the fileFormat parameter ('GeoTIFF' by default).

formatOptions parameter

Other configuration options are set with the formatOptions parameter, which should be a dictionary keyed by other format options, specific to each fileFormat as described below.

GeoTIFF

Cloud optimized GeoTIFF

To export a cloud optimized GeoTIFF, pass a JavaScript literal for formatOptions in which the cloudOptimized key is set to true. Continuing the previous example:

// Export a cloud-optimized GeoTIFF.
Export.image.toDrive({
  image: landsat,
  description: 'imageToCOGeoTiffExample',
  crs: projection.crs,
  crsTransform: projection.transform,
  region: geometry,
  fileFormat: 'GeoTIFF',
  formatOptions: {
    cloudOptimized: true
  }
});

import ee
import geemap.core as geemap

# Export a cloud-optimized GeoTIFF.
task = ee.batch.Export.image.toDrive(
    image=landsat,
    description='imageToCOGeoTiffExample',
    crs=projection['crs'],
    crsTransform=projection['transform'],
    region=geometry,
    fileFormat='GeoTIFF',
    formatOptions={'cloudOptimized': True},
)
task.start()

Cloud optimized GeoTIFFs can be reloaded from Cloud Storage into an Image. See the Image overview docs for details.

Nodata

Specify the GeoTIFF nodata value using the noData key within the formatOptions parameter. For example:

// Set a nodata value and replace masked pixels around the image edge with it.
var noDataVal = -9999;
landsat = landsat.unmask(noDataVal);

Export.image.toDrive({
  image: landsat,
  description: 'imageNoDataExample',
  crs: projection.crs,
  scale: 2000,  // large scale for minimal demo
  region: landsat.geometry(),  // full image bounds
  fileFormat: 'GeoTIFF',
  formatOptions: {
    noData: noDataVal,
  }
});

import ee
import geemap.core as geemap

# Set a nodata value and replace masked pixels around the image edge with it.
no_data_val = -9999
landsat = landsat.unmask(no_data_val)

task = ee.batch.Export.image.toDrive(
    image=landsat,
    description='imageNoDataExample',
    crs=projection['crs'],
    scale=2000,  # large scale for minimal demo
    region=landsat.geometry(),  # full image bounds
    fileFormat='GeoTIFF',
    formatOptions={'noData': no_data_val},
)
task.start()

Note that the nodata value should be inside the valid range for the image's PixelType. You can check PixelType by printing image metadata and looking at the first band's data_type property. You can also set the image'sPixelType by casting the data to a specific type using image methods toShort() or toInt(), for example.

TFRecord

See the TFRecord data format page.

maxPixels

The maxPixels parameter is intended to prevent very large exports from inadvertently being created. If the default value is too low for your intended output image, you can increase maxPixels. For example:

Export.image.toDrive({
  image: landsat,
  description: 'maxPixelsExample',
  crs: projection.crs,
  crsTransform: projection.transform,
  region: geometry,
  maxPixels: 1e9
});

Large file exports

If the output image is large, it will be exported as multiple files. If you are exporting to GeoTIFF(s), the image is split into tiles. The filename of each tile will be in the form baseFilename-yMin-xMin where xMin and yMin are the coordinates of each tile within the overall bounding box of the exported image.

If you are exporting to TFRecord, the files will be appended by -00000, -00001,... -0000N for N+1 files. Maintaining this order is important if you intend to perform inference on the files and upload the predictions back to Earth Engine as an image. See uploading images as TFRecord files for details.

Exporting images as they appear in the Code Editor

To export imagery as rendered on screen in Earth Engine, create visualization images as demonstrated in the Visualization images and the Compositing and Mosaicking sections. Since the Code Editor uses the 'EPSG:3857' CRS, specify a CRS of 'EPSG:3857' in the export to get an image in the same projection as that displayed in the Code Editor map. See the section on configuring image exports for details on specifying the resolution and coordinate system of the output.