This year (2017) I had again a contribution to the Salzburg-based GI_Forum Conference in the form of a multi-author full-paper about systematic flood assessment using Earth observation images. The GI_Forum is the English-speaking part of the AGIT Conference (AGIT-Website).
The title of my paper and presentation is “Automatic Ex-post Flood Assessment Using Long Time Series of Optical Earth Observation Images”, which I compiled earlier this year together with my co-authors Dirk Tiede, Lorenz Wendt and Andrea Baraldi.
The following picturesque photograph depicts the full spectacular scenery (image credit: Department of Geography, University of Tübingen):

If you are more interested in the science behind it, you can download the paper as open-access (Click here) or read the abstract:
Our study uses a dense temporal stack of 78 Landsat 8 images for surface water extraction using automatic Earth Observation (EO) image pre-processing, coupled with analyses over time for flood detection. The analysis is conducted with our IQ (ImageQuerying) system developed in-house, which allows ad-hoc executing of spatio-temporal queries against semantically enriched EO images. To facilitate high performance analyses, the data are stored as a spatio-temporal data cube in an array database. The analyses are automatically-translated database queries, which increase reproducibility, readability and comprehensibility for a human operator and can be conducted within just a few minutes. The specific analysis for this contribution is based on flood-extent mapping over different user-definable time spans. The results indicate areas that have been flooded at least once in the selected time span and are therefore prone to being flooded in future events. Additional spatial queries (e.g., for the indication of cloud cover) support the quality assessment of the flood analyses. We compared our result with a flood mask derived from a SAR (synthetic aperture radar) image of a single event in Somalia (Hiran province). Larger flooded areas overlap in both analyses, despite the non-synchronous acquisition times of the images. The results can be used as input for improved risk assessment and management of floods.

Categories Science, Big Data

Comments None

an interesting outcome of the end of the year 2016: Together with master students and Z_GIS employees I participated at the first Austrian Copernicus Hackathon in Vienna. The purpose of this hackathon was to create an application based on the freely available Sentinel 1, 2 or 3 data.

Our approach was to combine large-scale land use / land cover changes of agricultural fields with data regarding land ownership. We intent to detect the land use / land cover changes using satellite data and the land ownership with open (government) data and data based on crowd sourcing.

Happily, the jury awarded our concept and we won the first place! It was a head-to-head race with another team, and eventually the jury decided to let both of us win. Nevertheless, this was an important step for the whole team and myself. An outcome that I would never have expected.

The next step is to implement and improve our prototype with the help of our valuable partners, the EODC and Z_GIS. Visit our website on and stay tuned…

Categories Programming, Remote Sensing

Comments None

The semantic enrichment of Earth observation (EO) data is one major pillar for handling big Earth data. Therefore, we were very happy that the FFG-funded (Austrian Science Fund) SemEO project could be launched at Z_GIS in October 2016.

Find more information on the Z_GIS blog here

The picture above illustrates the semantic enrichment: Based on the raw data (top left corner), structure, local shape and color information will be (fully automated) extracted to achieve a scene classification (bottom right corner). Owner of this picture is Z_GIS.

My contribution to the project will be on the database and processing part: Given the semantically enriched data and the request for performing content-based searches and analyses, which is the best data model and data structure?
This gets especially complicated when geometrical, topological, temporal and semantic relationships need to be generated on-the-fly to support complex queries.

Categories Big Data, Science

Comments None

What a great honour to receive the 3rd prize of this year’s AGEO award for my work!

This is also a good opportunity to reflect on what has happened so far. Thanks for all major and minor contributions of those who have supported me and my work! Not only my current but also my past work with a constant change in topics and people. Everything and everyone was contributing and eventually lead to one path and I am truly grateful to them.

Ready for PhD and ready for Copernicus Sentinel 2 satellite big earth data.

Categories Study, Science

Comments None

What a great day for me! Today I was defending my master thesis at the Department of Geoinformatics – Z_GIS and passed the examination as the very last step for completion of my master’s study.

My master thesis was jointly supervised by experts in remote sensing (Dr. Dirk Tiede, Andrea Baraldi) and database technology (Univ.-Prof. Dr. Nikolaus Augsten). It focused on semantic Content-Based Image Retrieval (CBIR) and large-scale processing of remote sensing images using array databases. Due to the new generation of satellites carrying sensors with larger resolutions, the amount and volume of acquired images is increasing rapidly. Earth Observation and Earth Sciences are even today already facing big data issues. The full potential of remote sensing images is often not exploited since qualifying images cannot be found and retrieved efficiently in large image databases. Current implementations of remote sensing image retrieval systems provide only limited support for queries based on the image content. In most cases it is only possible to filter by textual metadata and criteria such as sensor type, date of acquisition and geographic extent etc. This leads to a significant amount of dark data (unused images) and hides information. The proposed solution is a scalable, web-based system which allows ad-hoc content-based (semantic) image retrieval and analyses of remote sensing images in large databases. It uses an array-DBMS (Database Management System) containing one or more automatically generated pre-classification maps (semantic layers) which are linked to the original images. By combining semantic layers with a graphical inference engine the system allows any content-based queries. The thesis describes the architecture of this system including a tailored concept for storing, retrieving, querying and analysing large amount of remote sensing images through space and time. Additionally, the thesis gives an overview of potentials and limitations of current array-DBMSs with special focus on the application of remote sensing imagery. The proposed system is implemented as an operational proof-of-concept which is called IQ (ImageQuerying).

Including some breaks I was in total 2 and a half years now in Salzburg and studying at the PLUS. Today, this major step is accomplished. I finished the master’s study with distinction – and new challenges are already waiting …

Categories Study

Comments None

← Older Newer →