2021-10-01, 10:00–10:30, Aconcagua
There’s a staggering amount of optical satellite data from the European Copernicus program. A large proportion of this data is not usable due to the intense cloud coverage many regions face during different periods of the year. Detecting clouds from optical satellite data is a crucial step for analysing the data. This presentation dives into the technical challenges and decisions we met during the design and implementation of a fully automated data pipeline for the United Nations Peace Operations to generate timely updated, cloud-free satellite imagery from high-resolution open data. Most peace operations are in tropical regions and prone to persistent cloud coverage throughout the year.
Open Data Cube (ODC) was utilised as the data platform of choice for harnessing the satellite data. ODC is an Open Source Geospatial Data Management and Analysis Software project that enables the use of satellite data in its broadest sense. ODC is a server-side software capable of processing and sharing immense amounts of satellite data.
The project was implemented as part of the United Nation’s Open GIS initiative (UN Open GIS).
With frequently updated, openly accessible satellite data and automatised processing it is possible to effectively monitor and evaluate different planetary changes. Cloud detection is inherently a part of analysing and working with optical satellite imagery such as Sentinel-2 data. In this presentation, we will discuss the data pipeline for automatising the data indexing, cloud detection, and image mosaicing process implemented with Open Data Cube (ODC). ODC is an open source server-side software capable of processing and sharing immense amounts of satellite data.
The project was implemented as part of the United Nation’s Open GIS initiative (UN Open GIS) and sought to enforce the Open Source GIS bundle that the UN operations are building principally for both peace-building and peace-keeping, but also for strengthening UN’s overall capabilities to use and leverage open source geospatial software as a global digital public good. At the same time, the agility and readiness of the UN cloud based infrastructure at UNGSC will be put into deal with operational challenges.
As a data source, we used optical satellite imagery derived from the European Copernicus Sentinel-2 mission. Sentinel-2 mission provides high-resolution satellite imagery and aims at monitoring different land conditions. The mission has a wide swath width (290 km) and a low revisit time (10 days at the equator with one satellite, and 5 days with 2 satellites). Sentinel-2 data is shared as open data and it is easy to access. The data pipeline was designed to use Sentinel-2 data from the “Registry of Open Data on AWS”. The adapted data processing and cloud detection workflow is also planned to be tested with other high resolution satellite imagery.
In the presentation, we will share insights on how the cloudless mosaic creation pipeline was designed, what cloud-based data processing infrastructure and which cloud detection algorithms we used, and which were some of the use cases that we tackled. Automated storing, processing, and sharing of optical satellite imagery is a widespread challenge as the world cumulates terabytes of data from different satellite data missions. This presentation contributes to solving this challenge by sharing insights on the use of ODC as part of the open source software stack for addressing these needs, in addition to the utilisation of Sentinel-2 as the main optical satellite imagery data source.
Vaaltola, Mikael (1)
Laine, Joona (1)
Sarkola, Pekka (1)
Pyykkönen, Santtu (1)
Thang, Cung (2)
Zhao, Yongping (2)
Stewart, Jonathan (2)
Ubukawa, Taro (2)
Obukhov, Timur (2)
Khan, Zeeshan (2)
(1) Gispo Ltd., Finland
(2) United Nations
Use cases & applicationsTopic –
Data collection, data sharing, data science, open data, big data, data exploitation platformsLevel –
2 - Basic. General basic knowledge is required.Language of the Presentation –
Mikael Vaaltola is a software developer specialized in developing geospatial software and modern software development methods. He has a strong knowledge of various open-source geospatial applications, including QGIS, PostGIS, and GeoServer. Vaaltola is also experienced with configuring and deploying software on different cloud platforms and operating systems.