基于ULS、TLS和超声测高仪的天然次生林中不同林冠层树高估测

【目的】应用不同数据源分析不同林冠层中探测提取树高的异同,探索适用于中国北方天然次生林树高估测的方法。【方法】以东北林业大学帽儿山实验林场中林施业区0.25 hm²样地为研究区域,基于无人机激光雷达(unmanned aerial vehicle laser scanning, ULS)、地基激光雷达(terrestrial laser scanning,TLS)和Vertex IV超声测高仪实测单木树高,根据冠层高度分布(canopy height distribution, CHD)对林冠层进行分层,对不同林冠层(上层和下层)、不同树木类型(针叶树和阔叶树)探测提取的树高进行对比与分析。【结果】由CHD计算得到的冠层分层阈值为8.5 m。树高的离群值大多产生在林冠上层,阔叶树比针叶树更容易产生离群值,ULS比TLS更容易产生离群值。在林冠上层,ULS比TLS估测树高的相对均方根误差(rRMSE)低2.56%,ULS提取针叶树树高的rRMSE比阔叶树低2.68%;在林冠下层,ULS仅能探测到少量树木,ULS比TLS探测提取树高的 rRMSE高6.31%,TLS提取针叶树树高的rRMSE比阔叶树低1.16%。【结论】针叶树的树高估测精度普遍高于阔叶树;当TLS和ULS均能对单木进行完全扫描时,具有准确提取树高的潜力;树高离群值多由冠型不规则或相互交叉的阔叶树产生,而大部分针叶树,由于具有规则的冠型,所以产生的离群值较少;基于CHD对林冠层进行划分能够较好地反映不同数据源估测树高的适用范围,具有一定的推广意义。

Tree height estimations for different forest canopies in natural secondary forests based on ULS,TLS and ultrasonic altimeter systems

【Objective】 A comparison of tree heights, estimated from different data sources for different forest canopies, was performed in natural secondary forests of Northern China to identify the optimal data source for different forest canopy layers. 【Method】 The study area, with a plot of 0.25 hm², was located in the Maoershan Forest Farm of the Northeast Forestry University. We conducted a performance comparison of terrestrial laser scanning (TLS), unmanned aerial vehicle laser scanning (ULS), and the Vertex IV ultrasonic instrument system in the estimation of individual tree heights. The forest canopy was stratified according to the canopy height distribution (CHD), and individual tree heights of different tree types (coniferous and broadleaved) were estimated and compared according to different canopy layers (overstory and understory). 【Result】 The threshold of canopy stratification calculated by CHD was 8.5 m. We also found that the outliers of tree heights mostly occurred in the overstory canopy, deciduous trees were more prone to outliers than coniferous trees, and ULS was more prone to outliers than TLS. Stratification of the forest canopy based on CHD can reflect the scope of tree height estimation using different data sources. In the overstory, the relative root mean square error (rRMSE) of tree height estimated by ULS was 2.56% lower than that estimated by TLS, while the rRMSE of coni-ferous tree height estimated by ULS was 2.68% lower than that of deciduous trees. ULS can only detect a small number of trees and had a higher rRMSE of 6.31% compared to TLS. Furthermore, the rRMSE of coniferous tree height estimated by TLS was 1.16% lower than that of deciduous trees. 【Conclusion】 For different canopy layers, the accuracy of conife-rous tree height was generally higher than that of deci-duous tree height. When both TLS and ULS can scan a single tree completely, both have the potential to accurately obtain the tree height. Most outliers of tree height were caused by deci-duous trees that had irregular or intertwined crowns. Most coniferous trees had few outliers in tree height due to their regular crown shapes. It is important to note that the stratification of the forest canopy based on CHD could reflect the scope of the applications of different data sources on the tree height estimation.