基于实例分割的高郁闭度林分单木树冠无人机遥感提取

目的 利用遥感影像获取高郁闭度林分树冠信息。 方法 试验了一种基于实例分割模型的无人机遥感影像单木树冠提取方法,选用7种残差网络用于模型的特征提取,逐一对不同郁闭度杉木纯林进行单木树冠提取。 结果 表明,7个实例分割模型对低郁闭度林分树冠分割的边界框AP值和掩膜AP平均值分别为55.89%、57.29%,林分东西冠幅、南北冠幅和树冠面积参数提取均方根误差平均值分别为0.161、0.179和0.341,平均预测决定系数R2分别为0.912、0.918和0.957;对高郁闭度林分树冠分割的边界框AP值和掩膜AP平均值分别为46.00%、44.45%,单木东西冠幅、南北冠幅和树冠面积参数提取均方根误差平均值分别为0.479、0.497和1.256,平均预测R2分别为0.806、0.762和0.936。 结论 各参数提取精度均优于传统调查精度,该方法能自动化、快速化、精准化获取树冠信息。     

Mapping of snow-damaged trees based on bitemporal airborne LiDAR data

The use of multitemporal LiDAR data in forest-monitoring applications has been so far largely unexplored. In this work, we aimed to develop and test a simple method for the detection of snow-induced canopy changes by employing bitemporal LiDAR data acquired in 2006–2010. Our study area was located in southern Finland (62°N, 24°E), where snow-induced damage occurred in 10 permanent Scots pine (Pinus sylvestris)-dominated plots in winter 2009–2010. For the detection of snow-damaged crowns, we developed a ?CHM method by contrasting bitemporal LiDAR canopy height models (CHMs) and analyzing the resulting difference image, using binary image operations to extract the damaged crowns. Furthermore, we examined the structural and spatial factors that could explain snow damage at the individual tree level. The ?CHM method developed is based on two threshold parameters, i.e., the required height difference in the contrasted CHMs and the minimum plausible area of damage. When testing the performance of ?CHM method, we found that the plot-level omission error rates were 19–75%, while the commission error rates were 0–21%. Furthermore, the relative estimation accuracy of the damaged crown projection area (DCPA) ranged from ?16.4 to 5.4%. The observed damage could be explained at tree level by stem tapering, relative tree size, and local stand density. To conclude, ?CHM method developed constitutes a potential tool for the monitoring of structural canopy changes in the dominant tree layer if dense multitemporal LiDAR data are available.     

无人机RGB影像中人工林单木位置的提取

从无人机RGB影像中提取单木位置时,由于树冠与非树冠植被的颜色相似,以及树冠之间存在粘连的问题,导致单木位置提取精度不高.针对这些问题,提出一种结合冠层高度模型(CHM)和形态学细化算法的人工林单木位置提取方法.首先根据无人机RGB影像生成数字正射影像(DOM)、数字高程模型(DEM)、数字表面模型(DSM),利用可见...     

基于无人机LiDAR的杉木树冠上部外轮廓模拟与可视化研究

目的 基于无人机激光雷达（LiDAR）点云数据提取杉木树冠上部结构参数（树冠顶点、树高、冠幅和上部冠长）,并进行树冠上部外轮廓模拟与可视化,为树种识别提供树木冠形特征。 方法 利用LASTools开源工具从激光雷达点云数据生成无孔洞的冠层高度模型,使用LiDAR360软件,采用局部最大值法检测树冠顶点,基于CHM种子点对点云进行单木分割,并在ArcGIS下手动选取杉木单株点云样本,用Python编程对“欠分割”样本进行单木纯化（之后全部编程方式自动化处理）;提取纯化后单株样本的树冠上部结构参数（树冠顶点、树高、冠幅和上部冠长）,再对单木点云按照一定高度间隔进行分层切片,使用宽度百分位数法提取单木树冠上部的相对着枝深度、枝条长度作为模型变量,以相对着枝深度分层分别建模与验证样本按照3倍标准差法剔除异常外轮廓点,选取二次多项式、幂函数和指数函数3个基础模型进行模型拟合与验证,最后采用最优拟合模型进行样地尺度的三维可视化。 结果 无人机激光雷达综合单木检测率为79.63%,结合实测参数与提取结果进行相关分析,树高线性回归R2为0.890 5,冠幅线性回归R2为0.845 6;二次抛物线、幂函数和对数函数拟合R2分别为0.807 0、0.817 0、0.806 0,幂函数对杉木树冠上部外轮廓的拟合效果更优。 结论 在高林分密度条件下,单木点云的有效提取纯化对客观描绘树冠形状非常重要;基于无人机激光雷达拟合的杉木树冠上部外轮廓反映了杉木的树冠上部形态,可为杉木的树种识别提供参考。     

基于机载激光雷达冠层高度模型的小班区划

目的 探究如何有效利用机载激光雷达冠层高度模型（CHM）自动区划小班,提高小班区划工作效率。 方法 在高光谱影像树种信息的辅助下,使用机载激光雷达数据生成的CHM进行两种空间尺度的分割和优化来自动区划小班。先对1 m空间分辨率CHM数据进行过分割,再对降尺度处理并平滑后的5 m空间分辨率CHM数据进行欠分割,结合两种尺度分割结果并优化得到最终区划结果。将自动区划结果与人工区划小班、数字正射影像（DOM）屏幕勾绘小班以及主伐作业小班为三类参考小班对比,采用最终测量精度（UMA）准则的圆度（RO）,紧致度（CO）,形状指数（SI）,最小包络圆短半径（RE）,椭圆度（EF）和形状因子（P2A）8个指标,及自动区划小班与参考小班的交并比（IOU）指标,定量评价自动区划小班边界勾绘的准确程度。并利用样地实测数据和CHM数据计算自动区划结果平均胸径、平均树高和冠层平均高的可解释性方差,验证自动区划结果的内部一致性和外部差异性精度。 结果 自动区划结果与参考小班的UMA形状、面积等特征较接近,与人工区划小班最相近。自动区划小班与人工区划、屏幕勾绘、主伐作业小班交并比大于70%的比例分别为46%,37%,43%,交并比大于50%的比例分别为61%,54%,55%。自动区划结果平均胸径可解释性方差为97%,平均树高可解释性方差为98%,和人工区划小班相同,说明其内部一致性高且和相邻小班差异大。冠层平均高可解释性方差为84.81%,比人工区划小班提高了1.77%。 结论 利用两种空间尺度的CHM与高光谱树种分类图的分割和优化方法自动区划的小班在内部一致性及边界的精准度方面有明显优势,更符合小班边界处林木的分布情况,小班边界准确,且工作效率高,有助于森林的精细化管理。     

Spatial partitioning of biomass and diversity in a lowland Bolivian forest: Linking field and remote sensing measurements

Large-scale inventories of forest biomass and structure are necessary for both understanding carbon dynamics and conserving biodiversity. High-resolution satellite imagery is starting to enable structural analysis of tropical forests over large areas, but we lack an understanding of how tropical forest biomass links to remote sensing. We quantified the spatial distribution of biomass and tree species diversity over 4 ha in a Bolivian lowland moist tropical forest, and then linked our field measurements to high-resolution Quickbird satellite imagery. Our field measurements showed that emergent and canopy dominant trees, being those directly visible from nadir remote sensors, comprised the highest diversity of tree species, represented 86% of all tree species found in our study plots, and contained the majority of forest biomass. Emergent trees obscured 1–15 trees with trunk diameters (at 1.3 m, diameter at breast height (DBH)) ≥20 cm, thus hiding 30–50% of forest biomass from nadir viewing. Allometric equations were developed to link remotely visible crown features to stand parameters, showing that the maximum tree crown length explains 50–70% of the individual tree biomass. We then developed correction equations to derive aboveground forest biomass, basal area, and tree density from tree crowns visible to nadir satellites. We applied an automated tree crown delineation procedure to a high-resolution panchromatic Quickbird image of our study area, which showed promise for identification of forest biomass at community scales, but which also highlighted the difficulties of remotely sensing forest structure at the individual tree level.     

小光斑激光雷达数据估测森林树高研究进展

小光斑激光雷达可以同时获得森林的垂直及水平结构参数,因光斑直径较小,可以做到森林单木结构参数的准确估计,进而推广到样方甚至更大区域森林结构参数的估计,近年来在林业中得到广泛应用。文中主要从树高估计方面对小光斑激光雷达在林业中的应用进行研究,通过对先前类似文献进行归纳总结发现,在小光斑激光雷达估测森林树高方面仍存在着一些问题,从而限制了森林树高估测精度的提高,如点云分类算法、点云密度、森林郁闭度、单木的准确分割等,还对小光斑激光雷达估计森林树高中所存在的问题进行了概括,并提出了改进建议。     

基于FCM和分水岭算法的无人机影像中林分因子提取

【目的】研究高精度小型无人机获取林分调查因子方法,将林分调查因子在低空无人机影像上识别并提取出来,获取树高、冠径等测树因子,建立林分因子测量方法,实现经济、高效、快捷、精准的森林资源调查和监测,及时掌握森林资源及相关林分因子的时空变化特征。【方法】以东北林业大学城市林业示范基地樟子松人工林为研究对象,以多旋翼无人机影像为数据源,基于FCM聚类算法和分水岭分割算法以及形态学运算、阈值分割、图像平滑、灰度化、二值化等一系列数字图像处理技术,提取樟子松人工林林分因子。FCM聚类算法和阈值分割法用于提取树梢标记图像,分水岭分割算法对树梢标记图像进行迭代处理从而获得单木树冠分割图像,根据单木树冠分割结果提取单木特征进而计算各林分因子值。【结果】在林地提取中,根据影像的颜色特征绿度分割成功地将林地部分与非林地部分分离开来,确定单木树冠分割范围。在单木树冠分割中,阈值分割法和FCM聚类算法均可有效将树梢标记从林地图像中提取出来;将基于标记的分水岭分割算法用于单木树冠分割取得较好效果,大多数单木树冠被单独分割出来,但某些区域仍然存在一定的欠分割或过分割问题。在林分因子提取中,提取的林分因子包括林分郁闭度、林地面积、立木株数和平均冠幅,其中林分郁闭度的测量精度为96.67%,林地面积的测量精度为81.23%,立木株数和平均冠幅的测量精度与单木树冠分割中的树梢提取方法(阈值分割法和FCM聚类算法)及分水岭分割中的2个参数(形态学腐蚀的结构元素大小和中值滤波的窗口大小)有关。针对2种树梢提取方法,分别进行参数组合试验,结果显示2种树梢提取方法使用适当参数组合所得各林分因子测量精度均在80%以上,平均测量精度均在90%以上,其中阈值分割法的最高平均测量精度为94.49%,FCM聚类算法的最高平均测量精度为93.17%。【结论】利用无人机拍摄的人工林影像进行森林资源调查,将先进的计算机科学技术和无人机技术应用到林业领域中,可有效提高森林资源调查的效率和精度。本研究提出的林分因子提取方法适用于高郁闭度林分,测量精度满足实际需求。     

基于无人机可见光遥感的单木树高提取方法研究

利用目前流行的高分辨率可见光无人机遥感影像生成树木冠层高度模型,采用分水岭分割算法提取单木树高的研究具有重要理论和实践意义。以位于云南省富民县的天然云南松纯林为研究对象,通过大疆Phantom 4 Pro无人机获取低空可见光遥感影像,利用Pix4D Mapper对无人机影像进行预处理及三维重建,生成三维点云,利用LiDAR360处理三维点云,构建DSM,DEM并生成CHM;采用分水岭分割算法对不同郁闭度条件下获得的CHM进行单木分割及树高提取,对提取结果进行精度评价。结果表明:分水岭分割算法能够准确分割CHM,利用无人机可见光遥感影像进行单木树高提取是可行的;将基于无人机可见光影像提取的树高值与野外实地调查得到的树高值进行对比,R²为0.893,RMSE为1.23m,估测精度为87.58%;同时,林分郁闭度会对单木树高估测产生影响,根据不同郁闭度条件下提取的3组样木树高与实地测量树高的决定系数(R²)分别是0.857,0.939和0.921,RMSE分别为1.450,1.097,0.896m,在低郁闭度林分内树高估测的精度显著高于高郁闭度林分。     

Crown plasticity and neighborhood interactions of European beech (Fagus sylvatica L.) in an old-growth forest

Competition for canopy space is a process of major importance in forest dynamics. Although virgin and old-growth European beech (Fagus sylvatica L.) forests in Europe have been studied for many years, there are to date no studies of individual-tree crown plasticity and the way this is influenced by local neighborhood interactions in these forests. In this study, we analyzed crown plasticity and local neighborhood interactions of individual trees in the upper canopy of the old-growth beech forests of Serrahn, northeast Germany. In a 2.8-ha sample plot, we measured crown radii of all upper canopy trees and analyzed the direction and extent of crown asymmetry. Size, relative position, and distance of neighboring trees were used to construct vectors of neighborhood asymmetry within different distances from target trees. The crowns of beech trees showed strong morphological plasticity. Mean absolute and relative displacement of crown centers from the stem base were 1.95 m and 0.37, respectively. Circular–circular rank correlation coefficients between the direction of crown displacement and the direction of neighborhood pressure showed that trees strongly positioned their crowns away from local neighbors. Highest correlation coefficients were obtained when basal area and relative position of neighboring trees within a radial distance of 12 m were considered. Clark and Evans index and Ripley’s K-function showed that crowns were more regularly distributed than stems. Projected canopy cover was about 10% higher than canopy cover with simulated circular crowns. We conclude that the crowns of older beech trees have a high ability to plastically respond to changes in the local canopy conditions, enabling very effective exploitation of canopy space.     

基于树高和树冠因子的立木材积与地上生物量相容模型研究

【目的】无人机机载激光雷达能够准确地测定单木、林分乃至大尺度森林结构参数(树高和树冠因子)。为应用无人机激光雷达技术准确估测森林蓄积量、生物量和碳储量提供计量依据和技术支撑。【方法】以150株实测马尾松生物量样本数据为研究对象,采用非线性回归估计方法和度量误差联立方程组方法,分析立木材积和地上生物量与树高、树冠因子的相关性,并在此基础上研究建立基于树高和树冠因子的立木材积与地上生物量相容模型。【结果】单株材积和地上生物量与树高因子的相关性最为紧密,其次才是树冠因子;基于树高和冠幅因子的二元材积和地上生物量模型预估精度较高,达到92%以上,再考虑冠长因子的三元模型预估精度改进不大;基于树高和冠幅因子的二元立木材积与地上生物量相容模型估计效果更好,相对于一元相容模型系统而言,二元相容模型拟合效果有较大幅度提高,预估精度达到92%以上。【结论】采用度量误差联立方程组方法可以有效解决基于树高和树冠因子的立木材积与地上生物量相容问题,并且预估精度达到92%以上,所建二元立木材积与地上生物量相容模型可为应用激光雷达技术反演森林蓄积量和生物量提供计量依据。     

Leaf traits in relation to crown development, light interception and growth of elite families of loblolly and slash pine

Crown architecture and size influence leaf area distribution within tree crowns and have large effects on the light environment in forest canopies. The use of selected genotypes in combination with silvicultural treatments that optimize site conditions in forest plantations provide both a challenge and an opportunity to study the biological and environmental determinants of forest growth. We investigated tree growth, crown development and leaf traits of two elite families of loblolly pine (Pinus taeda L.) and one family of slash pine (P. elliottii Mill.) at canopy closure. Two contrasting silvicultural treatments -- repeated fertilization and control of competing vegetation (MI treatment), and a single fertilization and control of competing vegetation treatment (C treatment) -- were applied at two experimental sites in the West Gulf Coastal Plain in Texas and Louisiana. At a common tree size (diameter at breast height), loblolly pine trees had longer and wider crowns, and at the plot-level, intercepted a greater fraction of photosynthetic photon flux than slash pine trees. Leaf-level, light-saturated assimilation rates (A(max)) and both mass- and area-based leaf nitrogen (N) decreased, and specific leaf area (SLA) increased with increasing canopy depth. Leaf-trait gradients were steeper in crowns of loblolly pine trees than of slash pine trees for SLA and leaf N, but not for A(max). There were no species differences in A(max), except in mass-based photosynthesis in upper crowns, but the effect of silvicultural treatment on A(max) differed between sites. Across all crown positions, A(max) was correlated with leaf N, but the relationship differed between sites and treatments. Observed patterns of variation in leaf properties within crowns reflected acclimation to developing light gradients in stands with closing canopies. Tree growth was not directly related to A(max), but there was a strong correlation between tree growth and plot-level light interception in both species. Growth efficiency was unaffected by silvicultural treatment. Thus, when coupled with leaf area and light interception at the crown and canopy levels, A(max) provides insight into family and silvicultural effects on tree growth.     

Site quality assessment of a Pinus radiata plantation in Victoria,Australia, using LiDAR technology

The aim of site quality assessment of Pinus radiata plantations is to determine the quality and productivity of the growing stock at different sites. It provides a useful indication of the site productivity to assist in the allocation of optimum thinning and fertiliser regimes and the scheduling of silvicultural operations. The predominant stand height (PDH) at a specific reference age, also known as site index (SI), is often used for site quality assessment of Pinus radiata plantations in Australia, as it is closely correlated with site productivity. However, measuring PDH in the field can be a time- and resource-consuming task. This paper proposes the use of light detection and ranging (LiDAR) data to estimate PDH for assessing the site quality of Pinus radiata. LiDAR provides highly accurate digital elevation and surface data that can be used to build a canopy height model (CHM). In this study, the state-of-the-art image segmentation technique, marker-controlled watershed segmentation, was employed for identifying locations of individual trees and estimating their heights from a CHM. Using an empirically derived SI equation, PDHs with reference age 11 years (SI₁₁) were estimated from the tallest trees identified in each forest stand, and were then used to determine the site quality class for each stand. The comparison of LiDAR-derived tree heights with field measurements produced an RMSE value of 0.42 m. The maximum horizontal distance between the field-measured locations of individual trees and the LiDAR-detected locations of their treetops was 1.87 m. Site quality classification was conducted in terms of 0.05 ha gridded plots, which revealed more detailed spatial variations of site quality across the study area than classification based on management plots. The study demonstrated that LiDAR provides an effective and accurate method for site quality classification of Pinus radiata.     

机载激光雷达和航空数码影像单木树高提取

用激光雷达(LiDAR)数据和航空数码影像相结合进行单木水平树高反演.对研究区的LiDAR点云数据进行滤波和分类,根据地形特点、地表植被状况以及其他地类的分布,采用Tin Filter滤波算法提取地面回波点和植被回波点.用面向对象的方法对高空间分辨率(25 cm)的航空数码影像进行单株木检测.通过多尺度、树冠模式的分割创建影像对象和类层次,用最邻近距离和成员函数法进行影像对象的分类,并基于分类结果进行再分割.对分割后的树冠多边形进行边缘优化,以准确识别单株木.将植被回波点和影像分割后得到的树冠多边形进行叠加,计算多边形内的LiDAR数据最大高程差值,与实测树高进行相关分析,建立单木树高估测回归方程,平均估测精度为74.89%.     

Competition and tree crowns: A neighborhood analysis of three boreal tree species

Competition for canopy space is a fundamental structuring feature of forest ecosystems and remains an enduring focus of research attention. We used a spatial neighborhood approach to quantify the influence of local competition on the size of individual tree crowns in north-central British Columbia, where forests are dominated by subalpine fir (Abies lasiocarpa), lodgepole pine (Pinus contorta) and interior spruce (Picea glauca × engelmanii). Using maximum likelihood methods, we quantified crown radius and length as functions of tree size and competition, estimated by the species identity and spatial arrangement of neighboring trees. Tree crown size depended on tree bole size in all species. Given low levels of competition, pine displayed the widest, shortest tree crowns compared to the relatively long and narrow crowns found in spruce and fir. Sensitivity to crowding by neighbors declined with increasing tree height in all but the pine crown radius model. Five of the six selected best models included separate competition coefficients for each neighboring tree species, evidence that species generally differ in their competitive effects on neighboring tree crowns. The selected crown radius model for lodgepole pine, a shade-intolerant species, treated all neighbors as equivalent competitors. In all species, competition from neighbors exerted an important influence on crown size. Per-capita effects of competition across different sizes and species of neighbors and target trees varied, but subalpine fir generally displayed the strongest competitive effects on neighbors. Results from this study provide evidence that species differ both in their response to competition and in their competitive influence on neighbors, factors that may contribute to maintaining coexistence.     

Vegetation and weather explain variation in crown damage within a large mixed-severity wildfire

The 2002 Biscuit Fire burned through more than 200,000 ha of mixed-conifer/evergreen hardwood forests in southwestern Oregon and northwestern California. The size of the fire and the diversity of conditions through which it burned provided an opportunity to analyze relationships between crown damage and vegetation type, recent fire history, geology, topography, and regional weather conditions on the day of burning. We measured pre- and post-fire vegetation cover and crown damage on 761 digital aerial photo-plots (6.25 ha) within the unmanaged portion of the burn and used random forest and regression tree models to relate patterns of damage to a suite of 20 predictor variables. Ninety-eight percent of plots experienced some level of crown damage, but only 10% experienced complete crown damage. The median level of total crown damage was 74%; median damage to conifer crowns was 52%. The most important predictors of total crown damage were the percentage of pre-fire shrub-stratum vegetation cover and average daily temperature. The most important predictors of conifer damage were average daily temperature and “burn period,” an index of fire weather and fire suppression effort. The median level of damage was 32% within large conifer cover and 62% within small conifer cover. Open tree canopies with high levels of shrub-stratum cover were associated with the highest levels of tree crown damage, while closed canopy forests with high levels of large conifer cover were associated with the lowest levels of tree crown damage. Patterns of damage were similar within the area that burned previously in the 1987 Silver Fire and edaphically similar areas without a recent history of fire. Low-productivity sites on ultramafic soils had 92% median crown damage compared to 59% on non-ultramafic sites; the proportion of conifer cover damaged was also higher on ultramafic sites. We conclude that weather and vegetation conditions — not topography — were the primary determinants of Biscuit Fire crown damage.     

Detection of individual trees and estimation of tree height using LiDAR data

For estimation of tree parameters at the single-tree level using light detection and ranging (LiDAR), detection and delineation of individual trees is an important starting point. This paper presents an approach for delineating individual trees and estimating tree heights using LiDAR in coniferous (Pinus koraiensis, Larix leptolepis) and deciduous (Quercus spp.) forests in South Korea. To detect tree tops, the extended maxima transformation of morphological image-analysis methods was applied to the digital canopy model (DCM). In order to monitor spurious local maxima in the DCM, which cause false tree tops, different h values in the extended maxima transformation were explored. For delineation of individual trees, watershed segmentation was applied to the distance-transformed image from the detected tree tops. The tree heights were extracted using the maximum value within the segmented crown boundary. Thereafter, individual tree data estimated by LiDAR were compared to the field measurement data under five categories (correct delineation, satisfied delineation, merged tree, split tree, and not found). In our study, P. koraiensis, L. leptolepis, and Quercus spp. had the best detection accuracies of 68.1% at h = 0.18, 86.7% at h = 0.12, and 67.4% at h = 0.02, respectively. The coefficients of determination for tree height estimation were 0.77, 0.80, and 0.74 for P. koraiensis, L. leptolepis, and Quercus spp., respectively.     

油松—白桦混交林种间关系研究

在调查河北隆化县油松－白烨混交林的林木生物量（地上部生物量及根系生物量）、林下草灌木、林地枯落物的总量以及不同混交距离油松地上地下部分生长发育状态的基础上，研究了混交林的种间关系。结果表明，该混交林作为华北地区广泛分布的一种林型，林下草灌木繁茂，林地枯落物分解迅速，有利于土壤理化性质和林分营养状况的改善，种间关系处于基本协调状态。由于林龄及密度原因，混交林生物量略低于油松纯林，在混交林中白桦对油松的生长有一定的抑制作用，抑制作用因混交距离而变化。本文提出了合适的混交距离和混交方式。