Application of Terrestrial Close-Range Photogrammetry for Estimating Stem Volume of Tree Species in Subtropical Forest in Okinawa, Japan
Application of Terrestrial Close-Range Photogrammetry for Estimating Stem Volume of Tree Species in Subtropical Forest in Okinawa, Japan
复制标题
DOI：
10.15684/formath.004
复制
发表时间：
2019-04
期刊：
影响因子：
0.5
通讯作者：
Y. Tamaki;M. Konoshima
中科院分区：
文献类型：
--
作者：
Y. Tamaki;M. Konoshima
为了解决大数据量的问题，提高数据质量，提高我们对冲绳树干体积估算的理解，迫切需要一种非破坏性、实用和高效的原位树木测量方法。这个问题变得很重要，因为两个与林业有关的政策问题在冲绳得到了很多关注。这两个问题是引入碳铈（cid:12）阳离子计划和将冲绳岛北部指定为世界自然遗产。在这项研究中，我们的注意力转向了最近出现的一种低成本方法，用于估算茎体积，同时对场地的环境影响最小，即陆地近距离摄影测量。我们探讨了该方法的有效性并评估了其局限性。为了检验地面摄影测量的准确性，我们将摄影测量估计的茎体积与使用3D磁运动跟踪器直接测量计算的茎体积进行了比较。估计的体积误差为0.0053 m3，估计的表面积误差为0.0714 m2。为了探索摄影测量的效率，我们扩展了我们的方法，包括在一个地块内同时重建多个树干的3D图像。利用一种最新的图像处理软件，我们成功地在一个地块内重建了11个茎的三维模型，并计算了每个茎的体积。估算的体积和表面积误差分别为0.0128 m3和0.1940 m2。
: 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 certi(cid:12)cation 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 m 3 root mean square error (RMSE) and the estimated surface was 0.0714 m 2 . 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 m 3 and 0.1940 m 2 , respectively.