Accurate delineation of individual tree crowns in tropical forests from aerial RGB imagery using Mask R-CNN
Accurate delineation of individual tree crowns in tropical forests from aerial RGB imagery using Mask R-CNN
复制标题
使用 Mask R-CNN 从航空 RGB 图像准确描绘热带森林中的单个树冠
DOI：
10.1101/2022.07.10.499480
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Ball J
中科院分区：
文献类型：
--
作者：
Ball J
热带森林是全球碳循环的重要组成部分，也是三分之二陆地物种的家园。树冠上部的树木储存了大部分森林碳，特别容易受到干旱和风暴的影响。监测它们的生长和死亡率对于了解森林对气候变化的适应能力至关重要，但在传统的实地调查中，大型树木的代表性不足，因此估计受到很大的限制。航空照片提供光谱和纹理信息，以区分不同，复杂的热带树冠，可能打开大门，景观监测的大树。在这里，我们描述了一种新的深度卷积神经网络方法Detectree 2，它建立在Mask R-CNN计算机视觉框架的基础上，可以从机载RGB图像中识别单个树冠的不规则边缘。我们在马来西亚婆罗洲的三个地点和法属圭亚那的一个地点用3,800个手动描绘的树冠训练和评估了这个模型。作为该工具的一个示例应用，我们将描绘与对四个地点的重复激光雷达调查相结合，以估计上层树冠树木的生长和死亡率。Detectree 2在14平方公里的航空图像中描绘了65，000棵上层树冠树木。的技能的自动方法在划定看不见的测试树木是好的（F1得分= 0.64），并为最高类别的树木是优秀的（F1得分= 0.74）。从以前的实地研究预测，我们发现，生长速度下降与树高和高大的树木有较高的死亡率比中等大小的树木。此外，法属圭亚那的树木比婆罗洲的树木有更高的生长率和死亡率。我们的方法表明，深度学习方法可以在广泛访问的RGB图像中自动分割树木。这个工具（作为一个开源Python包提供）在森林生态和保护方面有许多潜在的应用，从估计碳储量到监测森林物候和恢复。我们展示了它在跟踪树冠上部树木的生长和死亡率方面的用途，其规模远远大于现场数据。Python软件包可在https://github.com/PatBall1/Detectree2上安装
Tropical forests are a major component of the global carbon cycle and home to two-thirds of terrestrial species. Upper-canopy trees store the majority of forest carbon and can be particularly vulnerable to drought events and storms. Monitoring their growth and mortality is essential to understanding forest resilience to climate change, but large trees are underrepresented in traditional field surveys, so estimates are poorly constrained.Aerial photographs provide spectral and textural information to discriminate between tree crowns in diverse, complex tropical canopies, potentially opening the door to landscape monitoring of large trees. Here we describe a new deep convolutional neural network method,Detectree2, which builds on the Mask R-CNN computer vision framework to recognise the irregular edges of individual tree crowns from airborne RGB imagery. We trained and evaluated this model with 3,800 manually delineated tree crowns at three sites in Malaysian Borneo and one site in French Guiana. As an example application of this tool, we combined the delineations with repeat lidar surveys of the four sites to estimate the growth and mortality of upper-canopy trees.Detectree2delineated 65,000 upper-canopy trees across 14 km2of aerial images. The skill of the automatic method in delineating unseen test trees was good (F1score = 0.64) and for the tallest category of trees was excellent (F1score = 0.74). As predicted from previous field studies, we found that growth rate declined with tree height and tall trees had higher mortality rates than intermediate-size trees. In addition, trees in French Guiana had higher growth and mortality rates than those in Borneo.Our approach demonstrates that deep learning methods can automatically segment trees in widely accessible RGB imagery. This tool (provided as an open source Python package) has many potential applications in forest ecology and conservation, from estimating carbon stocks to monitoring forest phenology and restoration. We demonstrate its use in tracking the growth and mortality rates of upper-canopy trees at scales much larger than those achievable with field data.Python package available to install at https://github.com/PatBall1/Detectree2
登录
查看更多内容
使用 (R)-BINOL-6,6-二羧酸合成新型手性 Zn-MOF 并使用填充其的 HPLC 柱分离对映体
DOI：
--
发表时间：
2015
期刊：
影响因子：
--
作者：
T. Yano;M. Goto;N. Tamai;H. Matsuki;大坪 泰洋・田中 耕一
通讯作者：
大坪 泰洋・田中 耕一
影响因子：
3
作者：
J. Kellner;L. Albert;John T Burley;K. Cushman
通讯作者：
K. Cushman
影响因子：
13.5
作者：
Coomes, David A.;Dalponte, Michele;Qie, Lan
通讯作者：
Qie, Lan
在卫星遥感分辨率下绘制热带森林功能变化取决于关键特征
影响因子：
7.9
作者：
E. Ordway;G. Asner;D. Burslem;S. Lewis;R. Nilus;R. Martin;Michael J. O’Brien;O. Phillips;L. Qie;N. Vaughn;P. Moorcroft
通讯作者：
P. Moorcroft
透过树木看森林：树级测量如何帮助了解森林动态。
影响因子：
9.4
作者：
P. Zuidema;P. van der Sleen
通讯作者：
P. van der Sleen