Epsilon Archive for Student Projects

Abstract

Light Detection and Ranging (LiDAR) is a sensor type used to collect data in several forestry processes, which can increase efficiency and productivity. The aim of this study was to compare the accuracy of a LiDAR-based decision support system (DSS) tree diameter at breast height (DBH) measuring functionality at a plot level with reference measurements from a caliper. The function of the DSS is to assist forest harvester operators in thinning operations, utilizing software-analyzed output from the LiDAR in real-time.

Data were collected by mounting the DSS on a backpack and walking around the borders of the study plots. Results show that for mean diameter, the system had a relative root mean square error (RMSE) of 3.9%, for the quadratic mean diameter (DQ), 4.3%, and for basal area weighted mean diameter (DGV), a relative RMSE of 6.1%. For basal area, the relative RMSE was 6.9%, and the relative RMSE for the number of stems per plot was 5.8%.

The standard deviation (SD) ranged between 9.6% and 38.6% and showed a significant difference between the measurement methods. The aggregated diameter distribution for all study plots shows that the lowest and highest ends of the distribution are underrepresented in DSS measurements, while values around the mean of the distribution are overrepresented. A well-calibrated system for measuring DBH can provide enhanced operator support with increased accuracy in real-time to reach predefined thinning goals.