Challenges of Image-Based Crowd-Sourcing for Situation Awareness in Disaster Management
- Cite this paper as:
- Moreau G., Servières M., Normand JM., Magnin M. (2014) Challenges of Image-Based Crowd-Sourcing for Situation Awareness in Disaster Management. In: Teodorescu HN., Kirschenbaum A., Cojocaru S., Bruderlein C. (eds) Improving Disaster Resilience and Mitigation - IT Means and Tools. NATO Science for Peace and Security Series C: Environmental Security. Springer, Dordrecht
One of the main issues for authorities in disaster management is to build clear situation awareness that is consistent and accurate in both space and time. Authorities usually have Geographical Information Systems (GIS) for reference data. One current trend of GIS is the use of crowd sourcing, i.e. gather information from the public to build very large databases that would be very costly otherwise. The OpenStreetMap project is one famous example of this. One can easily imagine that, provided some communications networks are available, photographs shot with smartphones could help to build knowledge about the disaster scene. Advantages would be that costly sensor networks would be less necessary, more objectivity would come from pictures than from the public interpretation and that a huge amount of data could be collected very fast. Yet this raises a number of challenges that we will discuss in this paper: the first is geo-referencing of pictures, i.e. locating the pictures received with respect to the current GIS. Even if GPS are more and more common on smartphones, not all pictures will have GPS coordinates and GPS accuracy might not be enough. The second one is the need for a spatio-temporal data model to store and retrieve redundant and uncertain data, the third one is the need for new visualization techniques given the amount of data that will be stored. Finally, legal and ethical issues are raised by the use of massive images.