Adaptable region growing tool for fast mapping of natural disasters

¿ Forest fires, floods and windstorms are the major natural disasters in Europe. They cause loss of human lives, damage on housing, infrastructure, agriculture and forestry. We have used variety of aerial and satellite remote sensing (RS) data to map the most affected areas after several major disasters. Since the time for analysis is always limited to few days, semi-automatic tool was developed to speed the mapping of the affected areas. The tool is using region growing technique with the automatic capability to refine initial seed of the grown region. It is adaptable to various spatial and spectral resolutions. Although the knowledge of a trained operator is crucial and indispensable, this technique considerably speeds the mapping process. The behavior of the tool is fully under control of the operator. The parameters which govern the mapping algorithm can be adjusted according to the analyzed RS data and spectral response of the affected areas. The tool is written in IDL programming language and implemented under ENVI GUI interface. It helped to capture disaster extents using various RS data during the mapping of forest fires in Greece (2007), floods in UK (2007), floods in Romania (2008) and in other cases. Disaster maps of the several affected regions available from the partner organizations were used for validation of our mapping results with satisfactory spatial agreement.

KUCERA Jan; 

2011-08-29

International Center for Remote Sensing of Environment (ICRSE)

JRC54968

https://publications.jrc.ec.europa.eu/repository/handle/JRC54968,