Detection of forest damage by time series analysis of multispectral optical data
In work package 4, existing methods and processing lines for detection of forest damage will be adapted to the user needs and the specifics of forest types and structures of German forests and further developed. The services to be developed relate to recording of forest damage in "near real time", i.e. directly after availability of cloud-free Sentinel-2 data, and to generation of annual forest damage maps. The methods are based on time series analyses of Sentinel 2 data and refer to the detection of changes by means of special filtering methods as well as tointerpretation or classification of the cause of damage.
Involved Project Partner:
Scroll to top