1st Stage

class rsdtlib.Retrieve(starttime, endtime, aoi, shconfig)

Class for downloading observations from Sentinel Hub (https://www.sentinel-hub.com/).

Parameters:

  • starttime (datetime.datetime) – Starting time of the time series to retrieve

  • endtime (datetime.datetime) – End time of the time series to retrieve

  • aoi (str) – Area of Interest (path to shape file)

Example:

Define a download of observations in the time frame 01-01-2017 to 01-07-2017 from Sentinel Hub for the Area of Interest (AoI) defined in the shape file ostrava.shp. An account on Sentinel Hub is required for which the shconfig provides the access credentials.

import rsdtlib
from dateutil.parser import isoparse
from sentinelhub import SHConfig

# Credentials to access Sentinel Hub. See instructions:
# https://github.com/sentinel-hub/eo-learn/blob/master/examples/README.md
shconfig = SHConfig()
shconfig.instance_id = "<YOUR INSTANCE ID>"
shconfig.sh_client_id = "<YOUR CLIENT ID>"
shconfig.sh_client_secret = "<YOUR CLIENT SECRET>"

# Just a small area in Ostrava/CZ
my_aoi = "./ostrava.shp"

retrieve = rsdtlib.Retrieve(
                starttime=isoparse("20170101T000000"),
                endtime=isoparse("20170701T000000"),
                aoi=my_aoi,
                shconfig=shconfig)
get_images(datacollection, dst_path, maxcc=1.0)

Download the observations for the specified remote sensing type datacollection and maximum cloud coverage maxcc (if applicable). The observations are stored on the filesystem at dst_path.

Parameters:

  • datacollection (sentinelhub.DataCollection) – Data collection to download

  • dst_path (str) – Path on filesystem to store observations at

  • maxcc (float) – Maximum cloud coverage (default = 1.0)

Returns:

Number of retrieved observations

Return type:

int

Example:

Start the download of the AoI and time frame specified in the object retrieve. Sentinel Hub’s data collection SENTINEL1_IW_ASC is specified to retrieve Sentinel 1 observations in ascending orbit direction. The retrieved observations are stored in path dst_s1_asc.

from sentinelhub import DataCollection

dst_s1_asc = "<PATH FOR OBSERVATIONS>"

num_down = retrieve.get_images(
                    datacollection=DataCollection.SENTINEL1_IW_ASC,
                    dst_path=dst_s1_asc)
class rsdtlib.Convert(dst_path, aoi, normalize=255.0)

Class for converting GeoTIFF files to EOPatches.

Parameters:

  • dst_path (str) – Path on filesystem to store the converted observations at

  • aoi (str) – Area of Interest (path to shape file)

  • normalize (float) – Divisor to use for normalization (e.g., 255.0 for 8 bit unsigned integer types)

Example:

Define a conversion of observations for the Area of Interest (AoI) defined in the shape file ostrava.shp. The converted observations are stored in in path dst_path.

import rsdtlib

dst_path = "<PATH FOR OBSERVATIONS>"

# Just a small area in Ostrava/CZ
my_aoi = "./ostrava.shp"

convert = rsdtlib.Convert(
                dst_path=dst_path,
                aoi=my_aoi)
process(root_path, bands_tiff, mask_tiff, timestamp)

Start the conversion process. Merge the GeoTIFF with the bands bands_tiff and the GeoTIFF containing a mask mask_tiff from the same directory root_path. Annotate the result with a timestamp timestmap and store it as a single EOPatch.

Parameters:

  • root_path (str) – Root directory of all GeoTIFF files

  • bands_tiff (str) – Filename of the GeoTIFF containing the bands

  • mask_tiff (str) – Filename of the GeoTIFF containing the mask

  • timestamp (datetime.datetime) – Timestamp of the converted observation

Returns:

EOPatch object

Return type:

eolearn.core.EOPatch

Example:

Convert observations as specified in the object convert. This is called for every single observation provided as two GeoTIFF files from the root_path. One contains the contents (bands_tiff) and one the data mask (mask_tiff). A timestamp is assigned via timestamp.

root_path = "<ROOT OF GEOTIFFS>"

sample = convert.process(
            root_path=root_path
            bands_tiff="19931010T090051.TIF",   # LS5 TM 7 bands
            mask_tiff="19931010T090051_QA.TIF", # LS5 TM QA band
            timestamp="19931010T090051")