AI opens the door to single-tree-based forestry

New research is making it easier to identify both individual trees and their different parts using laser scanning data.

An increasing amount of forest is being digitalised through high-resolution laser scanning—using aircraft, helicopters, drones, or ground-based personnel. Modern laser scanners can map forest structure in a completely new way, opening new possibilities for managing forests at the single-tree level.

A laser scanner emits pulses of infrared light into the surroundings, and the reflected light points are captured by the scanner. Together, they form a point cloud that describes the surface of the ground and the vegetation, including the trees in the forest. This information can then be used to describe forest characteristics, tree species composition, and potential forest management measures.

To help develop a new approach to forest management, researchers from NIBIO and ETH Zurich in Switzerland have explored innovative ways to identify individual trees using different lidar data collection methods. By applying machine learning and neural networks, they have developed new data-driven models that provide users with much more detailed descriptions of individual trees in the forest.

The EU’s new forest strategy and the drive to enhance biodiversity by 2030 highlight the need for a shift towards forest management that places greater emphasis on multifunctional forest services. This new technology enables a transition from stand-level forest management to single-tree-based forest management.