Application of Terrestrial Close-Range Photogrammetry for Estimating Stem Volume of Tree Species in Subtropical Forest in Okinawa, Japan

A nondestructive, practical and efficient in-situ tree measurement approach is desperately needed to address the problem of large data volume, enhancing data quality and improving our understanding of stem volume estimation in Okinawa. This issue has become important because of two forestry related...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:FORMATH 2019/04/06, Vol.18, pp.004
Hauptverfasser: Tamaki, Yuki, Konoshima, Masashi
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:A nondestructive, practical and efficient in-situ tree measurement approach is desperately needed to address the problem of large data volume, enhancing data quality and improving our understanding of stem volume estimation in Okinawa. This issue has become important because of two forestry related policy issues that have received lots of attention in Okinawa. The two issues are the introduction of carbon certification program and the nomination of northern Okinawa Island as a World Natural Heritage site. In this study, our attention was turned to a more recent emerging low cost approach for estimating stem volume with minimum environmental impact on the site, that is, the terrestrial close-range photogrammetry. We explored the usefulness and evaluated the limitations of the approach. For examination of the accuracy of terrestrial photogrammetry, we compared the stem volume estimated from photogrammetry with the stem volume computed based on direct measurement using 3D magnetic motion tracker. The error of the estimated volume was 0.0053 m3 root mean square error (RMSE) and the estimated surface was 0.0714 m2. In order to explore the efficiency of the photogrammetry, we extended our approach to include the simultaneous reconstruction of 3D images of multiple tree stems within a plot. Using one of the latest image-processing software, we successfully reconstructed 3D models of 11 stems within a plot and computed stem volume for each. The error of the estimated volume and surface were 0.0128 m3 and 0.1940 m2, respectively.
ISSN:2188-5729
2188-5729
DOI:10.15684/formath.004