|Natural Disasters
Natural Disasters2016-10-19T16:43:33+02:00

Project Description

Natural Disasters

can be observed and analysed on a large scale world-wide via remote sensing.

Not only for the insurance industry but also for administrations, authorities and the civil defence it can be important to know about the extent of a natural disaster. Damage can thus be estimated and protection measures can be planned. Also, past events can be analysed using historical data.

Example: Storm damage in coniferous forests

Using satellite data of the years 2004 and 2005, coniferous forest was classified to identify the areas affected by storm damage via change detection.

Forest map before storm (yellow, left), loss of coniferous forest due to the storm (red, right)

It was shown that about 1.53% of the coniferous forests with a total area of 31 732 km² (damaged area 496 km²) were damaged by the storm.

Coniferous forest was classified with an accuracy of 96%.

Percentaged fraction of storm damaged forest.
Only forest area greater equal 1ha was considered

The results are taken from a study that was conducted together with SwissRe Germany AG and was supported by the European Space Agency (ESA): “Exploiting opportunities for integration of EO-based services with the reinsurance industry“.