Silvilaser 2019 - Poster Presentations »
Terrestrial Laser Scanning: an alternative for collecting dendrometric data in Atlantic secondary forests
Monitoring forest resources requires accurate, consistent and well-defined measurement techniques. This accuracy is a result from both the techniques and the data measuring instruments. In this way, the aim of this study was to compare dendrometric variables estimates achieved by conventional dendrometry (method I) and by terrestrial laser scanning (TLS) (method II) in a Atlantic secondary forests located in Brazil’s southeastern. A total of 29 trees belonging to 10 different species were selected, whose diameter at breast height, (DBH) were equal to or greater than 5 cm. Those trees had their DBH, total and commercial heights measured using applied methods. In method II, the TLS data were processed in Riscan Pro and 3D Forest softwares. The Pearson's linear correlation coefficient (ρ) was calculated to compare DBH, commercial and total height estimates obtaining from both methods. Among the dendrometric variables considered, the DBH exhibit very close values for aforesaid methods, followed by the total height. These variables estimated had a very strong positive correlation (ρ ≥ 0.9). However, it was noted that total height values derived by the method II presented a tendency of overestimation when they were compared to the method I. The commercial height estimates showed moderate positive correlation (0.5 ≤ ρ ≤ 0.7) between methods. The multiple scans approach, which guarantees complete stem circumference coverage, can explain the very strong correlation found among the DBH estimates. Another factor that contributed to the high correlation between DBH estimates, is the height of the LiDAR (close to 1.30 m), which prevents view angle errors and consequently points cloud distortions. On the other hand, factors such as crowns overlapping, high view angle and/or manual data filtering may explain the total height overestimation by method II. Regarding to the commercial height, the moderate correlation among values can be elucidated simply by the fact that in a native environment there are subjectivities in first branch definition and its detection, which can be different among observers and observations. Ultimately, we concluded that the TLS may be an alternative for dendrometric variables measurement highlighting the opportunity to automatize the process and to measure at the millimeter scale.