Drones and high-resolution optical sensors
Dr. Aravind Harikumar’s paper (PDF, Ensminger Lab) was published in IEEE Xplore: Fuzzy Approach to Individual Tree Crown Delineation in Uav Based Photogrammetric Multispectral Data
A. Harikumar, P. D'Odorico and I. Ensminger, "A Fuzzy Approach to Individual Tree Crown Delineation in Uav Based Photogrammetric Multispectral Data," IGARSS 2020 - 2020 IEEE International Geoscience and Remote Sensing Symposium, Waikoloa, HI, USA, 2020, pp. 4132-4135, doi: 10.1109/IGARSS39084.2020.9324303.
The use of drones and high-resolution optical sensors can provide detailed information on trees and entire forest canopies at sub-centimeter resolution and at unprecedented speed. The information obtained through this approach is critical for understanding forest health, ecosystem function, as well as for forest conservation and tree breeding. Nevertheless, a challenge for the wider use of this approach is a very simple problem: How can one distinguish one tree from another tree when looking at thousands of trees and a dense forest canopy from above? The precise localization and delineation of an individual tree crown is critical. Without this information it is impossible to map the biological and structural information contained in each data pixel collected by the drone sensor accurately to a single tree. This paper presents a fully automatic machine learning approach for the accurate localization and delineation of thousands of individual tree crowns in a forest with a complex overlapping canopy. The tool described is useful for a wide range of applications in basic research and for the forestry sector.
