基于数字化植物表型平台（D3P）的田间小麦冠层光截获算法开发

冠层光截获能力是反映作物品种间差异的重要功能性状，高通量表型冠层光截获对提高作物改良效率具有重要意义。本研究以小麦为研究目标，利用数字化植物表型平台（D3P）模拟生成了100种冠层结构不同的小麦品种在5个生育期的三维冠层场景，记录了从原始冠层结构中提取的绿色叶面积指数（GAI）、平均倾角（AIA）和散射光截获率（FIPAR_dif）信息作为真实值,进一步利用上述三维小麦场景开展了虚拟的激光雷达（LiDAR）模拟实验，生成了对应的三维点云数据。基于模拟的点云数据提取了其高度分位数特征（H）和绿色分数特征（GF）。最后，利用人工神经网络（ANN）算法分别构建了从不同LiDAR点云特征（H、GF和H+GF）输入到FIPAR_dif、GAI和AIA的反演模型。结果表明，对于GAI、AIA和FIPAR_dif，预测精度从高到低对应的点云特征输入为GF+H > H > GF。由此可见，H特征对提高目标表型特性的估算精度起到了重要作用。输入GF + H特征，在中等测量噪音（10%）情况下，FIPAR_dif和GAI的估算均获得了满意精度，R²分别为0.95和0.98，而AIA的估算精度（R²=0.20）还有待进一步提升。本研究基于D3P模拟数据开展，算法的实际表现还有待通过田间数据进一步验证。尽管如此，本研究验证了D3P协助表型算法开发的能力，展示了高通量LiDAR数据在估算田间冠层光截获和冠层结构方面的较高潜力。     

基于无人机LiDAR的天然林与人工林林隙提取

为研究主动遥感进行森林地物分类和林隙提取的效果,分别在天然林和人工林中比较了无人机激光雷达(Li DAR)数据的阈值法、逐像元法和面向对象法3种方法的分类精度和适用性。选取天然林(黑龙江省哈尔滨市帽儿山林场)和人工林(内蒙古自治区赤峰市旺业甸林场)两处试验区,应用阈值法、逐像元法和面向对象法3种方法,对两个试验区采集的无人机Li DAR数据进行林隙、非林隙、其他类型划分。研究结果表明,面向对象法在天然林和人工林试验区中的分类精度和Kappa系数均最高,天然林为82. 43%、0. 73,人工林为91. 74%、0. 88;逐像元法次之,天然林为76. 62%、0. 64,人工林为78. 68%、0. 68;阈值法的分类精度和Kappa系数差异较大,在天然林中的精度极低,为50. 54%、0. 27,人工林的精度较高,为79. 12%、0. 69。面向对象法和逐像元法在天然林和人工林普遍适用,均可以达到理想的分类精度和Kappa系数。阈值法在天然林的精度较低,更适合于人工林的分类,即林分高度趋于一致,且建筑、道路等其他类型干扰较少的区域。天然林的最佳分类方法为面向对象法,人工林的最佳分类方法为阈值法。     

基于DOM及LiDAR的多尺度分割与面向对象林隙分类

为研究分割尺度对航空正射影像(DOM)与LiDAR数据协同面向对象林隙分割与分类的影响,以东北典型的天然次生林帽儿山实验林场东林施业区为试验区,对DOM与LiDAR数据进行多尺度分割与面向对象林隙分类。分割过程中,采用基于DOM分割、基于LiDAR数据分割、DOMLiDAR协同分割3种分割方案。每种分割方案采用10种尺度。在每种尺度应用两种数据提取的光谱和高度两种特征,采用支持向量机分类器(SVM)进行林隙分类。研究结果表明:3种分割与分类方案分类精度随尺度的增大整体呈现下降的趋势,与ED3(Modified)趋势相反。基于LiDAR数据在尺度参数10获得了最优分割结果。在所有尺度上(10~100),基于LiDAR数据分割与分类精度高于其他两种数据源的分类精度,相比单独使用DOM优势更加明显。基于LiDAR数据分割与分类方案在尺度参数10时获得了最高分类精度(Kappa系数为80%)。3种分割与分类方案最优尺度的分类精度显著高于其他尺度分类精度。分割尺度对面向对象林隙分类结果有重要影响。     

Influence of tree crown characteristics on the local PM10 distribution inside an urban street canyon in Antwerp (Belgium): A model and experimental approach

Apart from influencing the amount of leaf-deposited particles, tree crown morphology will influence the local distribution of atmospheric particles. Nevertheless, tree crowns are often represented very rudimentary in three-dimensional air quality models. Therefore, the influence of tree crown representation on the local ambient PM₁₀ concentration and resulting leaf-deposited PM₁₀ mass was evaluated, using the three-dimensional computational fluid dynamics (CFD) model ENVI-met^® and ground-based LiDAR imaging. The modelled leaf-deposited PM₁₀ mass was compared to gravimetric results within three different particle size fractions (0.2–3, 3–10 and >10 μm), obtained at 20 locations within the tree crown. Modelling of the LiDAR-derived tree crown resulted in altered atmospheric PM₁₀ concentrations in the vicinity of the tree crown. Although this model study was limited to a single tree and model configuration, our results demonstrate that improving tree crown characteristics (shape, dimensions and LAD) affects the resulting local PM₁₀ distribution in ENVI-met. An accurate tree crown representation seems, therefore, of great importance when aiming at modelling the local PM distribution.     

基于机载LiDAR数据的玉米叶面积指数反演

以机载LiDAR离散点云数据为数据源,基于植被冠层孔隙率与叶面积指数的关系,提出一种反演大田玉米叶面积指数的方法。对反演LAI和实测LAI进行对比分析,结果表明:基于Axelsson改进的不规则三角格网加密方法可以将地面点和非地面点分开,结合高分辨率影像能够提取出玉米冠层点云;基于孔隙率反演LAI,尼尔逊参数的选择对结果影响很大,利用扫描天顶角模拟尼尔逊参数,LAI反演结果接近于真实情况。利用机载LiDAR点云数据能精确地反演大田玉米LAI,该研究方法适用于中等高度的农作物,可以扩展到甜菜、甘蔗等其他中等高度农作物。     

运用机载LiDAR数据对橡胶林地上生物量估测的采样尺度效应分析

气候变化下,森林生物量遥感监测是当前研究的热点,机载LiDAR作为重要的遥感信息源,其采样大小对生物量估测精度有着一定的影响。以机载LiDAR数据为信息源,以44块30m×30m的方形橡胶林实测样地数据为基础,对机载激光雷达数据进行不同尺寸采样(共21个采样尺寸,边长从10m至30m,间隔为1m),提取不同采样尺寸下的激光雷达参数,并与橡胶林地上生物量建立PLSR模型,就机载激光雷达采样大小对橡胶林地上生物量估测精度的影响进行研究。研究表明:当采样尺寸小于18m时,估测精度随着采样尺寸的增大而增大;而当采样尺寸大于18m时,估测精度随着采样尺寸的增大而减小,进而趋于平缓。结果虽然呈现出一定的规律性,但是差异并不是很明显。当采样尺寸为18m时估测效果最佳,模型决定系数(R^2)为0.718,均方根误差(RMSE)为17.830 t/hm^2;交叉验证精度P和RMSEcv分别为82.741%和18.874t/hm^2。相较于实际样地(30m)尺寸下的估测结果,18m采样尺寸下的R^2提高了1.989%,RMSEcv降低了2.611%。因此,生物量的估测精度受机载激光雷达数据采样尺寸大小的影响,在生物量估测过程中需结合研究对象和研究区的实际情况对采样尺寸进行选择,从而提高生物量估测精度。     

Estimation of urban tree canopy cover using random point sampling and remote sensing methods

Trees play an important role in urban areas by improving air quality, mitigating urban heat islands, reducing stormwater runoff and providing biodiversity habitat. Accurate and up-to-date estimation of urban tree canopy cover (UTC) is a basic need for the management of green spaces in cities, providing a metric from which variation can be understood, change monitored and areas prioritised. Random point sampling methods, such as i-Tree canopy, provide a cheap and quick estimation of UTC for a large area. Remote sensing methods using airborne Light Detection And Ranging (LiDAR) and multi-spectral images produce accurate UTC maps, although greater processing time and technical skills are required. In this paper, random point sampling and remote sensing methods are used to estimate UTC in Williamstown, a suburb of Melbourne, Australia. High resolution multi-spectral satellite images fused with LiDAR data with pixel-level accuracy are employed to produce the UTC map. The UTC is also estimated by categorising random points (a) automatically using the LiDAR derived UTC map and (b) manually using Google Maps and i-Tree canopy software. There was a minimum 1% difference between UTC estimated from the map derived from remotely sensed data and only 1000 random points automatically categorised by that same map, indicating the level of error associated with a random sampling approach. The difference between UTC estimated by remote sensing and manually categorised random point sampling varied in range of 4.5% using a confidence level of 95%. As monitoring of urban forest canopy becomes an increasing priority, the uncertainties associated with different UTC estimates should be considered when tracking change or comparing different areas using different methods.     

Economic utility of 3D remote sensing data for estimation of site index in Nordic commercial forest inventories: a comparison of airborne laser scanning,digital aerial photogrammetry and conventional practices

ABSTRACT

Forest productivity is a crucial variable in forest planning, usually expressed as site index (SI). In Nordic commercial forest inventories, SI is commonly estimated by a combination of aerial image interpretation, field assessment and information obtained from previous inventories. Airborne laser scanning (ALS) and digital aerial photogrammetry (DAP) data can alternatively be used for SI estimation, however the economic utilities of the inventory methods have not been compared. We compared seven methods of SI estimation in a cost-plus-loss analysis, by which we added the expected economic losses due to sub-optimal treatment decisions to the inventory costs. The methods comprised direct and indirect estimation from combinations of ALS, DAP and stand register data, and manual interpretation from aerial imagery supported by field assessment and information from previous inventories (conventional practices). The choice of method had great impact on both the accuracy and the economic value of the produced estimates. Direct methods using bitemporal ALS and DAP data gave the best accuracy and the smallest total cost. DAP was a suitable and low-cost data source for SI estimation. Estimation from single-date ALS and DAP data and age obtained from the stand register provided practical alternatives when applied to even-aged stands.     

Predicting individual stem volumes of sugi (Cryptomeria japonica D. Don) plantations in mountainous areas using small-footprint airborne LiDAR

This study investigated which predictor variables with respect to crown properties, derived from small-footprint airborne light detection and ranging (LiDAR) data, together with LiDAR-derived tree height, could be useful in regression models to predict individual stem volumes. Comparisons were also made of the sum of predicted stem volumes for LiDAR-detected trees using the best regression model with field-measured total stem volumes for all trees within stands. The study area was a 48-year-old sugi (Cryptomeria japonica D. Don) plantation in mountainous forest. The topographies of the three stands with different stand characteristics analyzed in this study were steep slope (mean slope ± SD; 37.6° ± 5.8°), gentle slope (15.6° ± 3.7°), and gentle yet rough terrain (16.8° ± 7.8°). In the regression analysis, field-measured stem volumes were regressed against each of the six LiDAR-derived predictor variables with respect to crown properties, such as crown area, volume, and form, together with LiDAR-derived tree height. The model with sunny crown mantle volume (SCV) had the smallest standard error of the estimate obtained from the regression model in each stand. The standard errors (m³) were 0.144, 0.171, and 0.181, corresponding to 23.9%, 21.0%, and 20.6% of the average field-measured stem volume for detected trees in each of these stands, respectively. Furthermore, the sum of the individual stem volumes, predicted by regression models with SCV for the detected trees, occupied 83%–91% of field-measured total stem volumes within each stand, although 69%–86% of the total number of trees were correctly detected by a segmentation procedure using LiDAR data.