Diameter at Breast Height (DBH) Estimation and Stem Cross-Section Shape Analysis of Eucalyptus Trees Using LiDAR Data after Noisy Removal

LiDAR data offer new possibilities for obtaining geometric parameters of forest areas, such as diameter at breast height (DBH), basal area, height, volume, biomass, and carbon stock. Terrestrial Laser Scanning (TLS) is highly accurate and can be used to obtain the shape of tree stem. In this paper, it is proposed a method for the automatic elimination of noisy points, followed by the classification of the cross-section shape of eucalyptus trees and the analysis of the relationship with the circular model, traditionally used to estimate DBH. Based on the proposed method, the DBH estimated from TLS data showed Root Mean Square Error (RMSE) of 0.7 cm for trees with a cross-section considered as circular and a RMSE of 3.7 cm for cross-sections considered as non-circular. The results showed that the shape of the cross-section is relevant for estimating parameters such as DBH, and that additional evaluations are needed for precise applications, such as volume estimation.