Eyewitness videos as an aid to crisis management


When a stream bursts its banks and floods cellars, many people record the event with their smartphones. Such eyewitness videos contain potentially interesting information for incident management. This project aims to identify relevant videos and filter their content for crisis management.

Portrait / project description (ongoing research project)

Humans are able to decide quickly and efficiently whether or not videos contain relevant information. In a crisis situation, however, there is not sufficient time to do so. The aim of this project is therefore to develop artificial intelligence methods to evaluate and filter videos. Selected videos will then be compared with known data and images from a region and correctly positioned and aligned. Depending on the content and settings, these steps will produce more or fewer false selections. Accordingly, the information will be visually prepared in such a way that the final selection and correction of the position can be performed simply and efficiently by crisis managers. This will ensure that the information is readily available and easy to use.


Online platforms contain many video recordings of a wide range of different incidents. In a stressful incident situation, however, crisis managers do not have the time to spend hours looking at videos in the hope of identifying relevant content. An automatic assessment and processing of eyewitness videos is intended to facilitate the use of information from different perspectives.


The objective is to support crisis managers with relevant information extracted from eyewitness videos by a combination of algorithms. The algorithms first have to assess the videos for relevance, and then analyse them with regard to content and correctly position and align them geographically. Because these processes entail uncertainties, the results are not always reliable. Nonetheless, appropriate visual processing ensures that the results can be efficiently and beneficially integrated into operational procedures.


The availability of current relevant information is crucial for crisis managers. They can draw on the information to make the right decisions fast and largely avoid costly wrong decisions with potential negative consequences for infrastructure as well as protection and rescue operations.

Original title

EVAC – Employing Video Analytics for Crisis Management

Project leaders

  • Prof. Susanne Bleisch, Institut Vermessung und Geoinformation, Hochschule für Architektur, Bau und Geomatik, Fachhochschule Nordwestschweiz
  • Prof. Jürgen Schmidhuber, Istituto Dalle Molle di Studi sull'Intelligenza Artificiale
  • Prof. Stephan Nebiker, Institut Vermessung und Geoinformation, Hochschule für Architektur, Bau und Geomatik, Fachhochschule Nordwestschweiz



Further information on this content


Prof. Susanne Bleisch Fachhochschule Nordwestschweiz
Hochschule für Architektur, Bau und Geomatik
Institut Vermessung und Geoinformation
Hofackerstrasse 30 4132 Muttenz

On this Subject