The European spruce bark beetle is considered to be the most critical disturbance agent in European forest ecosystems. To identify individual infested trees and thus minimize the risk of outbreak a random forest model was developed and validated using for shadow and forest clearing masked ortophotos, which were manually classified by foliage color, using known infestation areas. The model was able to discriminate between healthy and red-attacked trees on a pixel by pixel basis with an accuracy of 99%. This study highlights the potential use of such imagery as an effective tool for red-attack stage identification, as well as the necessity of increased temporal resolution of this data and more detailed in-situ data for future implementation and augmentation of this method in Latvia.
Lukass Roberts Kellijs