Reconstruction of tropical forest top of canopy using airborne structure from motion and terrestrial LiDAR technologies: an emperical analysis in the Guyanese reainforest

In recent years, terrestrial laser scanning (TLS) and unmanned aerial vehicles (UAVs) equipped with digital cameras, attracted much attention from the forestry community as possible tools for forest inventories. The present research fills a knowledge gap about the viability of using these technologies for modelling the top of canopy in the tropical forest. In an empirical study, with data acquired in the Guyanese t ropical forest, the differences between TLS derived top of canopy models (TCMs) and UAV structure from motion (SflvI) TCMs were assessed. This research determined that canopy gaps and crown architecture influences the way the TLS and UAV TCMs differ in respect to each other. Canopy gaps describe a clear pattern, the UAV TCM overestimating the canopy height in these areas (mean height differences between 0.19-13.83m). On the other hand, regarding crown architecture, a clear distinction between open crown and closed crown trees cannot be drawn, although there still can be observed definite differences between the open crown trees that have, on top, many leafless branches, and the rest of the trees. Secondly, this research shows that both TLS and UAV TCMs are sensitive to the moment of data collection. The TCMs rendered from UAV data acquired over the same area at different moments, are more similar (RMSE 0.11-0.63m for trees, and 0.14-3.05m for gaps) than TCMs rendered from TLS data acquired over the same area at different moments (RMSE 0.21-1.21m for trees, and 1.02-2.48m for gaps). This study offers the means for a more informed decision if choosing between TLS and UAV-Sflvl to assess tropical forest top of canopy by advancing the understandings on how these technologies model differently the top of the canopy and why, their ability to reproduce the same model over repeated surveying sessions, and also by providing insight about general considerations such as the area coverage, cost s, field-work time and processing requirements.

---

Download:

• 
  English-Full (12 MB)