Content uploaded by Matej Batic
Author content
All content in this area was uploaded by Matej Batic on Jun 23, 2022
Content may be subject to copyright.
Main objectives
GEM platform
Use cases
Address the challenge of continuous monitoring of large areas in a
sustainable and cost-effective way by establishing new and disrup-
tive Earth Observation data-exploitation model of Copernicus data
at a sustainable price.
Develop disruptive innovations with proprietary concept of Adjust-
able Data Cubes and integrate them with ML framework eo-learn.
Combine technological and Methodological innovations into a
unique continuous monitoring process based on seamless combi-
nation of multi-resolution data.
Validate the concept of continuous monitoring through the develop-
ment of ve specic use-cases and through their employment in a
6-month demonstration.
eo-learn is a collection of open-source Python packages that have been developed to
seamlessly access and process spatio-temporal image sequences acquired by any sat-
ellite eet in a timely and automatic manner. eo-learn is easy to use, it’s design mod-
ular, and encourages collaboration -- sharing and reusing of specic tasks in a typical
EO-value-extraction workows, such as cloud masking, image co-registration, feature
extraction, classication, etc. Everyone is free to use any of the available tasks and is
encouraged to improve the, develop new ones and share them with the rest of the com-
munity.
Built-up area
The goal of the built-up area use case is identifying
new built-up areas using the drill-down method. The
use-case by Sinergise exploits the Global Mosaic
ARD cube of Sentinel-2 data at 120 m resolution as
a starting point of the drill down mechanism. After
the fast detection of built-up areas at that resolu-
tion, the process runs on 10 m resolution to classi-
fy articial surfaces and to detect changes. At that
point, very high-resolution imagery is used to detect
buildings.
Map Making
TomTom map making support use-case will inte-
grate GEM Land Cover services to perform a fully
automated and repeatable global land cover map-
ping for small-mid scale and optimised land cover
map at large scale.
Crop-type identication
Meteoblue use case seeks to improve automat-
ic crop identication, validate the obtained results
through reliable ground truth data and enable auto-
matic crop growth stage identication. It relies on
Statistical API aDC approach to get the data for the
millions of elds across large AOIs and on TUM’s
expertise in ML. The use case supports operation-
al decisions when managing crops and the quanti-
tative monitoring of actual vs. planned or reported
land use (production forecast).
Conict pre-warning map use case
SatCen pilot explores the Climate-Security nex-
us. A Conict Pre-Warning (CPW) Map aims to
provide a new security product to support deci-
sion-making. It looks at correlations between glob-
al climate changes and environmental issues with
human activity behaviours, in support to guaran-
teeing the security of citizens.
eo-grow is an Earth observation framework for scaled-up processing in Python. The
eo-grow package takes care of running the workows at a large scale. In eo-grow an
EOWorkow based solution is wrapped in a pipeline object, which takes care of param-
etrization, logging, storage, multi-processing, EOPatch management and more. How-
ever, pipelines are not necessarily bound to EOWorkow execution and can be used for
other tasks such as training ML models.
Sentinel-Hub is a cloud API for satellite imagery that takes care of all the complexity of
handling satellite imagery archive and makes it available for end-users via easy-to-inte-
grate web services.
In GEM we exploit large-scale processing capabilities of Sentinel-Hub:
• Batch Processing API for creating pixel-based Adjustable Data Cubes
• Batch Statistical API for object-based Adjustable Data Cubes and Analysis Ready Data
• Bring Your Own Data for visualization and sharing of results
meteoblue is a weather service that offers access to environmental datasets like weath-
er simulations, ECMWF ERA5, gridded satellite information, soil properties and other
reanalysis datasets. Terabytes of data are accessible through the meteoblue Dataset
API. A unique feature of the meteoblue service is the option process data directly on
the service to perform expensive transformations and retrieve only the relevant infor-
mation from a dataset. The concept of transformations is exploited to offer environ-
mental data at satellite resolution in eo-learn.
The Sentinels Landsat
Collections
Commercial
Collections
DEM ENVISAT
MODIS Copernicus
services
Other Public
Collections
Bring Your
Own Data
Matej Batič1, Grega Milčinski1, Nicoletta Addimando2,
Francesco Perfetto3, Michele Lazzarini4, Marco Koerner5
1 Sinergise, 2 meteoblue, 3 TomTom, 4 European Union
Satellite Centre, 5 Technische Universität Munchen
continuous monitoring of large areas
in a sustainable and cost-effective way
Sentinel-2 L2A
120m Mosaic
© SentinelHub
Copernicus Sentinel-2
L2A data
© SentinelHub
SPOT
© AIRBUS DS 2020
Pleiades
© CNES 2020,
Distribution AIRBUS DS
Weather data Ground truth
data
EuroCrops Other (geo-int)
data
The project has received funding from European Union’s
Horizon 2020 Research and Innovation Programme” under
the Grant Agreement 101004112.
Open Imagery:
Copernicus, USGS
Machine
learning
API
OGC
Comercial Imagery:
Airbus, Planetscope,
Maxar
Bring Your Own Data:
Aerial Imagery, Other
Sensors
Web/Mobile Apps
GIS: Cloud and Desktop
Apps
Other Data: CORINE,
DEM
Sentinel-1 Sentinel-2
Automatically
computed
water mask