Find and download Landsat and Sentinel-2 data from the public Google Cloud
The script downloads the index.csv file listing all available Landsat or Sentinel-2 tiles, then searches the file for one scene that matches user parameters. Once found, it downloads the image files.
Small demo video: https://youtu.be/8zCs0nxl-rU
Developed with/for Python 2.7 and 3.3+ You may either install the package through pip:
pip install fels
or if using a conda environment, the following steps are recommended to create and install dependencies:
conda create --name fetchLSGC python=3
Switch to the new environment (source activate fetchLSGC
in Linux), and
install the gdal dependency from conda-forge
conda config --add channels conda-forge
conda install gdal
fels OLI_TIRS 2015-01-01 2015-06-30 -s 203031 -c 30 -o ~/LANDSAT --latest --outputcatalogs /tmp
fels S2 2016-10-01 2016-12-31 -s 44UPU -o ~/SENTINEL2 -l --outputcatalogs /tmp
You can also use GeoJSON geometry to perform a search:
fels OLI_TIRS 2015-01-01 2015-06-30 -g '{"type":"Polygon","coordinates":[[[-122.71,37.54],[-122.71,37.90],[-121.99,37.90],[-121.99,37.54],[-122.71,37.54]]]}' -c 30 -o ~/LANDSAT --latest --outputcatalogs /tmp
or you can use Well Known Text (WKT) geometry:
fels OLI_TIRS 2015-01-01 2015-06-30 -g 'POINT (-105.2705 40.015)' -c 30 -o ~/LANDSAT --latest --outputcatalogs /tmp
fels OLI_TIRS 2015-01-01 2015-06-30 -s 203031 -c 30 -o %TEMP%\LANDSAT --latest --outputcatalogs %TEMP%\LANDSAT
fels S2 2016-10-01 2016-12-31 -s 44UPU -o %TEMP%\SENTINEL2 -l --outputcatalogs %TEMP%\SENTINEL2
You can use the Python entrypoint fels.run_fels
in the same way as the fels
executable:
# these commands are equivalent
# CLI
import os
os.system(('fels OLI_TIRS 2015-01-01 2015-06-30 -c 30 -o . -g "POINT (-105.2705 40.015)"'
'--latest --outputcatalogs ~/data/fels/'))
os.system(('fels OLI_TIRS 2015-01-01 2015-06-30 -c 30 -o . -g \'{"type":"Point","coordinates":[-105.2705, 40.015]}\''
'--latest --outputcatalogs ~/data/fels/'))
# python
from fels import run_fels
urls = run_fels(None, 'OLI_TIRS', '2015-01-01', '2015-06-30', cloudcover=30, output='.',
geometry='POINT (-105.2705 40.015)',
latest=True, outputcatalogs=os.path.expanduser('~/data/fels/'))
print(urls)
# python with friendly aliases
from datetime import date
urls = run_fels(None, 'L8', date(2015, 1, 1), date(2015, 6, 30), cloudcover=30, output='.',
geometry={'type': 'Point', 'coordinates': [-105.2705, 40.015]},
latest=True, outputcatalogs=os.path.expanduser('~/data/fels/'))
print(urls)
and import other useful utilities like:
fels.safedir_to_datetime
fels.landsatdir_to_date
fels.convert_wkt_to_scene
Run the script with -h
switch for parameters:
usage: fels [-h] [-g GEOMETRY] [-c CLOUDCOVER] [-o OUTPUT] [-e EXCLUDEPARTIAL] [--latest]
[--noinspire] [--outputcatalogs OUTPUTCATALOGS] [--overwrite] [-l] [-d] [-r]
[scene] {TM,ETM,OLI_TIRS,S2} start_date end_date
Find and download Landsat and Sentinel-2 data from the public Google Cloud
positional arguments:
scene WRS2 coordinates for Landsat (ex 198030) or MGRS for S2 (ex 52SDG). Mutually
exclusive with --geometry
{TM,ETM,OLI_TIRS,S2} Which satellite are you looking for
start_date Start date, in format YYYY-MM-DD. Left-exclusive.
end_date End date, in format YYYY-MM-DD. Right-exclusive.
optional arguments:
-h, --help show this help message and exit
-g GEOMETRY, --geometry GEOMETRY
Geometry to run search. Must be valid GeoJSON `geometry` or Well Known Text (WKT).
This is only used if --scene is blank.
-i, --includeoverlap If -g is used, include scenes that overlap the geometry but do not
completely contain it
-c CLOUDCOVER, --cloudcover CLOUDCOVER
Set a limit to the cloud cover of the image
-o OUTPUT, --output OUTPUT
Where to download files
-e EXCLUDEPARTIAL, --excludepartial EXCLUDEPARTIAL
Exclude partial tiles - only for Sentinel-2
--latest Limit to the latest scene
--noinspire Do not rename output image folder to the title collected from the inspire.xml file
(only for S2 datasets)
--outputcatalogs OUTPUTCATALOGS
Where to download metadata catalog files
--overwrite Overwrite files if existing locally
-l, --list List available download urls and exit without downloading
-d, --dates List or return dates instead of download urls
-r, --reject_old For S2, skip redundant old-format (before Nov 2016) images
You can read more about the public google access to Landsat and Sentinel-2 data here: https://cloud.google.com/storage/docs/public-datasets/
Contributors (THANK YOU!):
- https://github.com/vascobnunes
- https://github.com/framioco
- https://github.com/bendv
- https://github.com/GreatEmerald
- https://github.com/dgketchum
- https://github.com/eriFelaP
- https://github.com/JuraBit
- https://github.com/matt-bernstein
- https://github.com/Erotemic
- https://github.com/banesullivan