Disaster events and major damage situations caused by floods, forest fires, extreme snow conditions or storms pose major challenges for civil protection in terms of (1) avail-ability of near real-time and large-scale information for disaster assessment and management, (2) analysis of the data in near real time, and (3) fusion of derived in-formation products for intuitive, transparent and focused decision support. The follow-ing problems arise from this:
P1: Lack of automated AI-based algorithms for analysis of innovative data sources to support situation awareness and assessment.
P2: Lack of availability of reliable AI methods for the fusion of information layers from the analysis of remote sensing and internet data in disaster management.
P3: Largely neglected consideration of legal, sociological and ethical framework con-ditions with regard to the data and AI methods used.
P4: Frequently non-demand-oriented research, which often does not generate usable results for end users.
P5: Insufficient consideration of the integration of project developments into existing, proven processes of disaster management.
The AIFER project addresses all of the above-mentioned problems to ensure civil securi-ty and better information availability during a natural disaster. Solutions:
S1: Research into explainable AI algorithms that automatically extract information from earth observation data (e.g. satellite data, aerial and drone images) and internet data (e.g. geo-social media, georeferenced news articles, Google Trends).
S2: Development of an AI-based algorithm for the fusion of information based on the analyses of earth observation and internet data.
S3: Legal, sociological and ethical issues are dealt with in detail, and examined and reflected with the help of an Ethical Board.
S4: End-user needs are collected and investigated in a structured way and integrated into technical research activities.
S5: Validation and integrability of the results are achieved through the strong involve-ment of end users and practical testing in a TRL4 test environment.
AIFER primarily addresses the disaster scenario of flooding with the analysis of two his-torical (“cold”) and a real-time (“warm”) use cases. The transferability to other disaster scenarios (forest fires, storms, extreme snowfall) will be validated.
ProjektleiterIn / Name und Institut/Unternehmen
Assoz.-Prof. Dr. Bernd Resch, Universität Salzburg, Fachbereich Geoinformatik – Z_GIS
Auflistung der weiteren Projekt- bzw. KooperationspartnerInnen
Österreichische Konsortialpartner
Universität Salzburg, Fachbereich Geoinformatik – Z_GIS
Institut für empirische Sozialforschung GmbH
Johanniter Österreich Ausbildung und Forschung gemeinnützige GmbH
Österreichisches Rotes Kreuz – Landesverband Salzburg
Spatial Services GmbH
Deutsche Konsortialpartner
Deutsches Zentrum für Luft- und Raumfahrt (DLR)
Universität Kassel
Bundesanstalt Technisches Hilfswerk (THW)
Bayerisches Rotes Kreuz
Disy Informationssysteme GmbH
Assoziierte Partner
Ärzte ohne Grenzen (MSF)
Rotes Kreuz Tirol
Rotes Kreuz Wien
Name / Institut oder Unternehmen
Universität Salzburg, Fachbereich Geoinformatik – Z_GIS
Kontakt
Assoz.-Prof. Dr. Bernd Resch
Universität Salzburg, Fachbereich Geoinformatik – Z_GIS
Schillerstraße 30
A-5020 Salzburg
Tel: +43-662-8044-7551
Fax: +43-662-8044-7560
E-mail: bernd.resch@sbg.ac.at
Website: zgis.at
Projekt-Website: https://giscience.zgis.at/aifer