In this work, the use of high-resolution satellite imagery for the detection and assessment of windfall damage is presented. We focused on analysing forest damage using a change detection method on multi-temporal optical satellite imagery. We analysed the changes in the Alpine valleys of Krma, Kot and Radovna in Triglav National Park, where a major windstorm occurred on 19 July 2023, leaving an enormous impact on forest land. Sentinel-2 (10 m spatial resolution) and PlanetScope (3 m spatial resolution) high-resolution optical satellite imagery images have been used, as well as digital orthophoto imagery from 2023, with manually delineated reference areas of damaged trees used for evaluating the scope of damage and accuracy control. We achieved an area accuracy of 77% for the Sentinel-2 and 81% for the PlanetScope imagery. We also developed a method to calculate damaged tree volume based on woodstock data and compared the volumes with in-situ, field-derived sanitary logging data provided by Slovenian Forest Service. Additional sample-based validation showed the reliability of the method, as the calculated volumes matched with in-situ data by an average of 7% for Sentinel-2 and 5% for PlanetScope imagery. That confirmed the method's applicability in forestry, as it allows fast and accurate estimation of the damage caused by windstorms.
|
