Abstract
The American beech (Fagus grandifolia) plays a key role throughout eastern North American forests. However, beech bark disease (BBD) causes widespread mortality of beech trees. We investigated whether imagery collected using an unpiloted aerial system (UAS) could differentiate beech tree health. Reference data were collected from 140 beech trees in New Hampshire and were visually classified as having “no/trace damage,” “moderate damage,” or “heavy damage.” Multispectral imagery was collected from a UAS, and 44 image features were derived for each beech crown. We used machine learning to identify the importance of each feature in distinguishing between the three health classes and found that textural features were more important overall. The classification accuracy was improved to over 70% when collapsing BBD classes into “healthy” or “unhealthy”. These results demonstrate that UAS imagery can identify areas afflicted by BBD, and be used to prioritize management applications.
Recommended Citation
Lopez, Isabelle; Fraser, Benjamin T.; and Congalton, Russell G.
(2023)
"Evaluating the use of Unpiloted Aerial Systems to detect and monitor beech bark disease in New England forests,"
The Geographical Bulletin: Vol. 64:
Iss.
2, Article 4.
Available at:
https://digitalcommons.kennesaw.edu/thegeographicalbulletin/vol64/iss2/4