Using eCognition Definiens for automated detection of snow avalanche deposits from very high resolution optical imagery - New developments

Abstract

The identification of snow avalanche deposits from high resolution optical satellite imagery had been the focus of the project "avalRS” which NGI, together with the Norwegian Computing Centre and Statens Veivesen, had carried out for the European Space Agency (2008-2011; e.g., Frauenfelder et al., 2011). The algorithms developed have produced variable results, often working well in certain situations and poorly in others. In 2011 using the object oriented image processing software eCognition, NGI developed two prototype algorithms on its own. The two algorithms were developed for (i) QuickBird satellite imagery, and (ii) Leica ADS-40 airborne imagery (cf. Lato and Frauenfelder, 2012).Aspart of the continuation of this research program, the algorithms developed in 2011 were published in the journal *Natural Hazards and Earth System Sciences* (Lato et al., 2012a) as well as presented at International conferences, e.g., at the "International Snow Science Workshop 2012” in Anchorage, Alaska (Lato et al., 2012b). Overall the developments have been accepted well within the community, the preliminary results demonstrate the possibility of numerous research and commercial applications. In parallel with the publication and presentation of the research results in 2012, new satellite images containing snow avalanche deposits were tested with the algorithms in eCognition. An overview of the data, the region it represents, as well as a discussion of the results is included in this document.