基于ICESatGLAS的云南省森林地上生物量反演

结合机载、星载激光雷达对GLAS(地球科学激光测高系统)光斑范围内的森林地上生物量进行估测,并利用MODIS植被产品以及MERIS土地覆盖产品进行了云南省森林地上生物量的连续制图。机载LiDAR扫描的260个训练样本用于构建星载GLAS的森林地上生物量估测模型,模型的决定系数(R2)为0.52,均方根误差(RMSE)为31Mg/ha。研究结果显示,云南省总森林地上生物量为12.72亿t,平均森林地上生物量为94Mg/ha。估测的森林地上生物量空间分布情况与实际情况相符,森林地上生物量总量与基于森林资源清查数据的估测结果相符,表明了利用机载LiDAR与星载ICESatGLAS结合进行大区域森林地上生物量估测的可靠性。     

GLAS星载激光雷达和Landsat/ETM+数据的森林生物量估算

基于大脚印激光雷达数据和野外观测数据,该文提出一种获取脚印点内森林生物量的新思路,并结合陆地卫星数据应用于长白山地区森林地上生物量估算。首先,基于3种森林类型(针叶林、阔叶林和针阔混交林),采用多元逐步回归方法建立激光雷达波形指数与脚印点内实测平均树高的回归模型,估算全部脚印点内的平均树高;然后根据脚印点内样方的野外观测数据(平均树高和平均胸径)以及它们与样方生物量的拟合方程估算没有野外调查数据对应的脚印点的生物量;最后对3种森林类型的脚印点森林生物量在各森林覆盖度条件下进行分层分区统计得到生物量等级图。验证比较遥感估算的生物量与野外调查数据推算的生物量,总体误差在0~30(t·hm~(-2))之间,均方根误差为14.66(t·hm~(-2))。     

Forest canopy height estimation using ICESat/GLAS data and error factor analysis in Hokkaido,Japan

Spaceborne light detection and ranging (LiDAR) enables us to obtain information about vertical forest structure directly, and it has often been used to measure forest canopy height or above-ground biomass. However, little attention has been given to comparisons of the accuracy of the different estimation methods of canopy height or to the evaluation of the error factors in canopy height estimation. In this study, we tested three methods of estimating canopy height using the Geoscience Laser Altimeter System (GLAS) onboard NASA’s Ice, Cloud, and land Elevation Satellite (ICESat), and evaluated several factors that affected accuracy. Our study areas were Tomakomai and Kushiro, two forested areas on Hokkaido in Japan. The accuracy of the canopy height estimates was verified by ground-based measurements. We also conducted a multivariate analysis using quantification theory type I (multiple-regression analysis of qualitative data) and identified the observation conditions that had a large influence on estimation accuracy. The method using the digital elevation model was the most accurate, with a root-mean-square error (RMSE) of 3.2 m. However, GLAS data with a low signal-to-noise ratio (⩽10.0) and that taken from September to October 2009 had to be excluded from the analysis because the estimation accuracy of canopy height was remarkably low. After these data were excluded, the multivariate analysis showed that surface slope had the greatest effect on estimation accuracy, and the accuracy dropped the most in steeply sloped areas. We developed a second model with two equations to estimate canopy height depending on the surface slope, which improved estimation accuracy (RMSE = 2.8 m). These results should prove useful and provide practical suggestions for estimating forest canopy height using spaceborne LiDAR.     

A novel approach to estimate canopy height using ICESat/GLAS data: A case study in the New Forest National Park,UK

The Geoscience Laser Altimeter System (GLAS) aboard Ice, Cloud and land Elevation Satellite (ICESat) is a spaceborne LiDAR sensor. It is the first LiDAR instrument which can digitize the backscattered waveform and offer near global coverage. Among others, scientific objectives of the mission include precise measurement of vegetation canopy heights. Existing approaches of waveform processing for canopy height estimation suggest Gaussian decomposition of the waveform which has the limitation to properly characterize significant peaks and results in discrepant information. Moreover, in most cases, Digital Terrain Models (DTMs) are required for canopy height estimation. This paper presents a new automated method of GLAS waveform processing for extracting vegetation canopy height in the absence of a DTM. Canopy heights retrieved from GLAS waveforms were validated with field measured heights. The newly proposed method was able to explain 79% of variation in canopy heights with an RMSE of 3.18 m, in the study area. The unexplained variation in canopy heights retrieved from GLAS data can be due to errors introduced by footprint eccentricity, decay of energy between emitted and received signals, uncertainty in the field measurements and limited number of sampled footprints.Results achieved with the newly proposed method were encouraging and demonstrated its potential of processing full-waveform LiDAR data for estimating forest canopy height. The study also had implications on future full-waveform spaceborne missions and their utility in vegetation studies.     

激光雷达森林参数反演研究进展

激光雷达通过发射激光能量和接收返回信号的方式,来获取高精度的森林空间结构和林下地形信息。全波形激光雷达通过记录返回信号的全部能量,得到亚米级植被垂直剖面;离散回波激光雷达记录的单个或多个回波,表示来自不同冠层的回波信号。星载激光雷达一般采用全波形或光子计数激光剖面系统,仅能获取卫星轨道下方的单波束或多波束数据,用于区域/全球范围的森林垂直结构及变化观测。机载激光雷达多采用离散回波或全波形激光扫描系统,能够获取飞行轨迹下方特定视场范围内的扫描数据,用于林分/区域范围的森林结构观测。地基激光雷达多采用离散回波激光扫描系统,获取以测站为中心的球形空间内扫描数据,用于单木/样地范围的森林结构观测。激光雷达单木因子估测方法可分为CHM单木法、NPC单木法和体元单木法3类。CHM单木法通过局部最大值识别树冠顶点,采用区域生长或图像分割算法识别树冠边界或树冠主方向,NPC单木法一般通过空间聚类或形态学算法识别单木,体元单木法在3维体元空间采用区域生长或空间聚类算法识别树冠。根据激光雷达冠层高度分布可以估测林分因子,冠层高度分布特征来自于离散点云或全波形。多时相激光雷达可用于森林生长量、生物量变化等监测,以及森林采伐、灾害等引起的结构变化监测。随着激光雷达技术的发展,它将在森林调查、生态环境建模等生产与科学研究领域中得到更为广泛的应用。     

吉林长白山森林冠顶高度激光雷达与MERSI联合反演

将激光雷达与光学遥感相结合进行区域林分冠顶高度联合反演,提出了大脚印激光雷达GLAS脚点波形数据处理和不同地形条件下的森林冠顶高度反演算法,并建立了区域尺度不同森林类型林分冠顶高度GLAS＋MERSI联合反演模型,制作了长白山地区森林冠顶高度图。     

机载激光雷达数据分析与反演青海云杉林结构信息

利用机载激光雷达点云数据,计算了9种度量指标,并将其分为冠层的高度指标、结构复杂度指标和覆盖度指标。利用高度指标和结构复杂度指标,结合大量实测单木结构与年龄估测数据,从样点和区域尺度分别分析了青海云杉林冠层垂直结构分布,分析得知实验区内主要以中龄林和成熟林为主,冠层垂直分布复杂程度偏低,高度分化程度一般。通过回归分析发现首次回波覆盖度指标FCI与实测的有效植被面积指数PAI_e有良好的相关性（R²=0.66）,在此基础上基于辐射传输模型反演了实验区内PAI_e的水平分布,且用实测数据验证发现反演的PAI_e略高于实测值（R²=0.67）,绝对平均误差为0.65。分析结果很好地反映了激光雷达在森林空间结构信息提取方面的应用潜力。     

高分二号异轨立体数据的森林高度提取

森林植被碳储量的空间分布格局及其动态变化是陆地生态系统碳收支核算的基础。作为森林地上生物量的重要指示因子,森林高度的精确估算是提高森林植被碳储量估算精度的关键。现有研究已证明,由专业星载摄影测量系统获取的立体观测数据可用于森林高度提取,但光学遥感数据最大的问题是受云雨等天气因素的影响严重。区域森林地上生物量产品的生产需要充分挖掘潜在数据源。国产高分二号卫星(GF-2)虽然不是为获取立体观测数据而设计的专业星载摄影测量系统,但其获取的图像空间分辨率可达0.8 m,且具备±35°的的侧摆能力,在重复观测区域可构成异轨立体观测。本文以分别获取于2015年6月20日和2016年7月19的GF-2数据作为立体像对,其标称轨道侧摆角分别为0.00118°和20.4984°,以激光雷达数据获取的林下地形(DEM)和森林高度(CHM)为参考,对利用GF-2立体观测数据进行森林高度提取进行了研究。通过对立体处理得到的摄影测量点云的栅格化得到DSM,以激光雷达数据提供的DEM作为林下地形,得到了GF-2的CHM。结果表明GF-2提取的CHM与激光雷达CHM空间分布格局较为一致,两者之间存在明显的相关性,像素对像素的线性相关性(R2)达到0.51,均方根误差(RMSE)为3.6 m。研究结果表明,在林下地形已知的情况下,GF-2立体观测数据可用于森林高度估算。     

激光雷达回波模型辅助的坡地森林冠层高度反演

大光斑激光雷达数据已广泛应用于森林冠层高度提取,但通常仅限于地形坡度小于20°的平缓地区。在地形坡度大于20°的陡峭山区,地形引起的波形展宽使得地面回波和植被回波信息混合在一起,给森林冠层高度提取带来巨大挑战。本文利用激光雷达回波模型和地形信息,提出了一种模型辅助的坡地森林冠层高度反演算法。该方法以激光雷达回波信号截止点为参考,定义了波形高度指数H50和H75,使用激光雷达回波模型与已知地形信息模拟裸地的激光雷达回波,将裸地回波信号截止点与森林激光雷达回波信号截止点对齐,利用裸地回波计算常用的波形相对高度指数RH50和RH75,对森林冠层高度进行反演。并与高斯波形分解法和波形参数法的反演结果进行了比较。研究结果表明:(1)利用所提取的波形指数RH50和RH75对胸高断面积加权平均高(Lorey’s height)进行了估算,在坡度小于20°时,高斯波形分解法、波形参数法和模型辅助法的估算结果与实测值线性拟合的相关系数(R2)分别为0.70,0.78和0.98,对应的均方根误差(RMSE)分别为2.90 m,2.48 m和0.60 m,模型辅助法略优于其他两种方法;(2)在坡度大于20°时,高斯波形分解法、波形参数法和模型辅助法的R2分别为0.14,0.28和0.97,相应的RMSE分别为4.93 m,4.53 m和0.81 m,模型辅助法明显优于其他两种方法;(3)在0°—40°时,模型辅助法对Lorey’s height估算结果与实测值的R2为0.97,RMSE为0.80 m。本研究提出的模型辅助法具有更好的地形适应性,在0°—40°的坡度范围内具备对坡地森林冠层高度反演的潜力。     

地基激光雷达的玉兰林冠层叶面积密度反演

叶面积密度LAD(Leaf Area Density)是表征冠层内部叶面积垂直分布的重要参数,其分布廓线的准确反演对研究植被碳氮循环、初级生产力和生物量估算等具有重要意义。本文在电子科技大学校内建立实验样区,利用地基激光雷达Leica Scan Station C10和数码相机获取玉兰林高分辨率3维激光点云数据和真彩色影像。利用监督分类将真彩色影像中枝干等非光合组织与叶片分离,再将像素分类信息映射给点云数据,从而提取叶片点云。通过点云数据体元化,并引入2维凸包算法确定垂直方向分层树冠边界,获取激光接触冠层的频率;随机选择不同高度的多个叶片,利用特征值法进行叶片平面拟合,估算出叶倾角,并结合天顶角估算叶倾角校正因子;最后基于体元的冠层分析VCP(Voxel-based Canopy Profiling)方法实现树林冠层LAD反演。结果表明体元化的叶片点云数据能准确确定树林冠层边界和统计接触频率实现LAD反演;反演的LAD变化走势与区域林木冠层叶片垂直分布相吻合,在冠层中下部随着高度的增加叶面积密度也随之增加,在4 m高度处达到最大值1 m2/m3,之后随着高度的增加叶面积密度逐渐降低。根据LAD计算得到的累积叶面积指数LAI为3.20 m2/m2,与LAI-2200实测的叶面积指数相比,相对误差为1.26%。     

基于遥感的区域尺度森林地上生物量估算研究

森林是陆地生态系统最大的碳库,精确估算森林生物量是陆地碳循环研究的关键。首先从机载LiDAR数据中提取高度和密度统计量,采用逐步回归模型进行典型样区生物量估算;然后利用机载LiDAR数据估算的生物量作为样本数据,与多光谱遥感数据Landsat8 OLI的波段反射率及植被指数建立回归模型,实现区域尺度森林地上生物量估算。实验结果显示,机载LiDAR数据估算的鼎湖山样区生物量与地面实测生物量的相关性R2达0.81,生物量RMSE为40.85 t/ha,说明机载LiDAR点云数据的高度和密度统计量与生物量存在较高的相关性。以机载LiDAR数据估算的生物量为样本数据,结合多光谱遥感数据Landsat8 OLI估算粤西北地区的森林地上生物量,精度验证结果为:R2为0.58,RMSE为36.9 t/ha;针叶林、阔叶林和针阔叶混交林等3种不同森林类型生物量的估算结果为:R2分别为0.51(n=251)、0.58(n=235)和0.56(n=241),生物量RMSE分别为24.1 t/ha、31.3 t/ha和29.9 t/ha,估算精度相差不大。总体上看,利用遥感数据可以开展区域尺度的森林地上生物量估算,为森林固碳监测提供有力的参考数据。     

A new 500-m resolution map of canopy height for Amazon forest using spaceborne LiDAR and cloud-free MODIS imagery

In the present study, we aimed to map canopy heights in the Brazilian Amazon mainly on the basis of spaceborne LiDAR and cloud-free MODIS imagery with a new method (the Self-Organizing Relationships method) for spatial modeling of the LiDAR footprint. To evaluate the general versatility, we compared the created canopy height map with two different canopy height estimates on the basis of our original field study plots (799 plots located in eight study sites) and a previously developed canopy height map. The compared canopy height estimates were obtained by: (1) a stem diameter at breast height (D) – tree height (H) relationship specific to each site on the basis of our original field study, (2) a previously developed D–H model involving environmental and structural factors as explanatory variables (Feldpausch et al., 2011), and (3) a previously developed canopy height map derived from the spaceborne LiDAR data with different spatial modeling method and explanatory variables (Simard et al., 2011). As a result, our canopy height map successfully detected a spatial distribution pattern in canopy height estimates based on our original field study data (r = 0.845, p = 8.31 × 10⁻³) though our canopy height map showed a poor correlation (r = 0.563, p = 0.146) with the canopy height estimate based on a previously developed model by Feldpausch et al. (2011). We also confirmed that the created canopy height map showed a similar pattern with the previously developed canopy height map by Simard et al. (2011). It was concluded that the use of the spaceborne LiDAR data provides a sufficient accuracy in estimating the canopy height at regional scale.     

High-resolution mapping of forest canopy height using machine learning by coupling ICESat-2 LiDAR with Sentinel-1, Sentinel-2 and Landsat-8 data

Forest canopy height is an important indicator of forest carbon storage, productivity, and biodiversity. The present study showed the first attempt to develop a machine-learning workflow to map the spatial pattern of the forest canopy height in a mountainous region in the northeast China by coupling the recently available canopy height (H_canopy) footprint product from ICESat-2 with the Sentinel-1 and Sentinel-2 satellite data. The ICESat-2 H_canopy was initially validated by the high-resolution canopy height from airborne LiDAR data at different spatial scales. Performance comparisons were conducted between two machine-learning models – deep learning (DL) model and random forest (RF) model, and between the Sentinel and Landsat-8 satellites. Results showed that the ICESat-2 H_canopy showed the highest correlation with the airborne LiDAR canopy height at a spatial scale of 250 m with a Pearson’s correlation coefficient (R) of 0.82 and a mean bias of -1.46 m, providing important evidence on the reliability of the ICESat-2 vegetation height product from the case in China’s forest. Both DL and RF models obtained satisfactory accuracy on the upscaling of ICESat-2 H_canopy assisted by Sentinel satellite co-variables with an R-value between the observed and predicted H_canopy equalling 0.78 and 0.68, respectively. Compared to Sentinel satellites, Landsat-8 showed relatively weaker performance in H_canopy prediction, suggesting that the addition of the backscattering coefficients from Sentinel-1 and the red-edge related variables from Sentinel-2 could positively contribute to the prediction of forest canopy height. To our knowledge, few studies have demonstrated large-scale vegetation height mapping in a resolution ≤ 250 m based on the newly available satellites (ICESat-2, Sentinel-1 and Sentinel-2) and DL regression model, particularly in the forest areas in China. Thus, the present work provided a timely and important supplementary to the applications of these new earth observation tools.     

An improved method for estimating forest canopy height using ICESat-GLAS full waveform data over sloping terrain: A case study in Changbai mountains,China

Light Detection And Ranging (LiDAR) has a unique capability for estimating forest canopy height, which has a direct relationship with, and can provide better understanding of the aboveground forest carbon storage. The full waveform data of the large-footprint LiDAR Geoscience Laser Altimeter System (GLAS) onboard the Ice, Cloud, and land Elevation Satellite (ICESat), combined with field measurements of forest canopy height, were employed to achieve improved estimates of forest canopy height over sloping terrain in the Changbai mountains region, China. With analyzing ground-truth experiments, the study proposed an improved model over Lefsky's model to predict maximum canopy height using the logarithmic transformation of waveform extent and elevation change as independent variables. While Lefsky's model explained 8–89% of maximum canopy height variation in the study area, the improved model explained 56–92% of variation within the 0–30° terrain slope category. The results reveal that the improved model can reduce the mixed effects caused by both sloping terrain and rough land surface, and make a significant improvement for accurately estimating maximum canopy height over sloping terrain.     

Integration of WorldView-2 and airborne LiDAR data for tree species level carbon stock mapping in Kayar Khola watershed,Nepal

Integration of WorldView-2 satellite image with small footprint airborne LiDAR data for estimation of tree carbon at species level has been investigated in tropical forests of Nepal. This research aims to quantify and map carbon stock for dominant tree species in Chitwan district of central Nepal. Object based image analysis and supervised nearest neighbor classification methods were deployed for tree canopy retrieval and species level classification respectively. Initially, six dominant tree species (Shorea robusta, Schima wallichii, Lagerstroemia parviflora, Terminalia tomentosa, Mallotus philippinensis and Semecarpus anacardium) were able to be identified and mapped through image classification. The result showed a 76% accuracy of segmentation and 1970.99 as best average separability. Tree canopy height model (CHM) was extracted based on LiDAR’s first and last return from an entire study area. On average, a significant correlation coefficient (r) between canopy projection area (CPA) and carbon; height and carbon; and CPA and height were obtained as 0.73, 0.76 and 0.63, respectively for correctly detected trees. Carbon stock model validation results showed regression models being able to explain up to 94%, 78%, 76%, 84% and 78% of variations in carbon estimation for the following tree species: S. robusta, L. parviflora, T. tomentosa, S. wallichii and others (combination of rest tree species).     

Individual tree crown delineation using localized contour tree method and airborne LiDAR data in coniferous forests

Individual tree crown delineation is of great importance for forest inventory and management. The increasing availability of high-resolution airborne light detection and ranging (LiDAR) data makes it possible to delineate the crown structure of individual trees and deduce their geometric properties with high accuracy. In this study, we developed an automated segmentation method that is able to fully utilize high-resolution LiDAR data for detecting, extracting, and characterizing individual tree crowns with a multitude of geometric and topological properties. The proposed approach captures topological structure of forest and quantifies topological relationships of tree crowns by using a graph theory-based localized contour tree method, and finally segments individual tree crowns by analogy of recognizing hills from a topographic map. This approach consists of five key technical components: (1) derivation of canopy height model from airborne LiDAR data; (2) generation of contours based on the canopy height model; (3) extraction of hierarchical structures of tree crowns using the localized contour tree method; (4) delineation of individual tree crowns by segmenting hierarchical crown structure; and (5) calculation of geometric and topological properties of individual trees. We applied our new method to the Medicine Bow National Forest in the southwest of Laramie, Wyoming and the HJ Andrews Experimental Forest in the central portion of the Cascade Range of Oregon, U.S. The results reveal that the overall accuracy of individual tree crown delineation for the two study areas achieved 94.21% and 75.07%, respectively. Our method holds great potential for segmenting individual tree crowns under various forest conditions. Furthermore, the geometric and topological attributes derived from our method provide comprehensive and essential information for forest management.     

Estimation of LAI and above-ground biomass in deciduous forests: Western Ghats of Karnataka, India

This study demonstrates the potentials of IRS P6 LISS-IV high-resolution multispectral sensor (IGFOV  6 m)-based estimation of biomass in the deciduous forests in the Western Ghats of Karnataka, India. Regression equations describing the relationship between IRS P6 LISS-IV data-based vegetation index (NDVI) and field measured leaf area index (ELAI) and estimated above-ground biomass (EAGB) were derived. Remote sensing (RS) data-based leaf area index (PLAI) image is generated using regression equation based on NDVI and ELAI (r² = 0.68, p ≤ 0.05). RS-based above-ground biomass (PAGB) image was generated based on regression equation developed between PLAI and EAGB (r² = 0.63, p ≤ 0.05). The mean value of estimated above-ground biomass and RS-based above-ground biomass in the study area are 280(±72.5) and 297.6(±55.2) Mg ha⁻¹, respectively. The regression models generated in the study between NDVI and LAI; LAI and biomass can also help in generating spatial biomass map using RS data alone. LISS-IV-based estimation of biophysical parameters can also be used for the validation of various coarse resolution satellite products derived from the ground-based measurements alone.     

An object-based approach for mapping forest structural types based on low-density LiDAR and multispectral imagery

Mapping forest structure variables provides important information for the estimation of forest biomass, carbon stocks, pasture suitability or for wildfire risk prevention and control. The optimization of the prediction models of these variables requires an adequate stratification of the forest landscape in order to create specific models for each structural type or strata. This paper aims to propose and validate the use of an object-oriented classification methodology based on low-density LiDAR data (0.5 m^?2) available at national level, WorldView-2 and Sentinel-2 multispectral imagery to categorize Mediterranean forests in generic structural types. After preprocessing the data sets, the area was segmented using a multiresolution algorithm, features describing 3D vertical structure were extracted from LiDAR data and spectral and texture features from satellite images. Objects were classified after feature selection in the following structural classes: grasslands, shrubs, forest (without shrubs), mixed forest (trees and shrubs) and dense young forest. Four classification algorithms (C4.5 decision trees, random forest, k-nearest neighbour and support vector machine) were evaluated using cross-validation techniques. The results show that the integration of low-density LiDAR and multispectral imagery provide a set of complementary features that improve the results (90.75% overall accuracy), and the object-oriented classification techniques are efficient for stratification of Mediterranean forest areas in structural- and fuel-related categories. Further work will be focused on the creation and validation of a different prediction model adapted to the various strata.     

机器学习算法在森林地上生物量估算中的应用

森林地上生物量是森林生产力的重要评价指标,对其进行高效监测对维持全球碳平衡和保护生态系统具有重要意义。本文首先基于冠层高度模型数据,通过分水岭分割算法得到单木冠幅边界;然后在单木冠幅范围内提取23个LiDAR变量,结合佩诺布斯科特试验森林的87组实测数据,利用随机森林和支持向量机建立森林地上生物量估算模型;最后对样地模型估算的结果进行了比较,讨论了预测结果及其精度。结果表明:本文选用的随机森林模型和支持向量机模型在估算森林地上生物量的应用中获得了较高的精度;并且,随机森林模型在基于机载雷达数据估测森林地上生物量中的估算精度更高,模型泛化能力更强,制图精度也更好,具有更好的适用性。     

3D segmentation of single trees exploiting full waveform LIDAR data

This paper highlights a novel segmentation approach for single trees from LIDAR data and compares the results acquired both from first/last pulse and full waveform data. In a first step, a conventional watershed-based segmentation procedure is set up, which robustly interpolates the canopy height model from the LIDAR data and identifies possible stem positions of the tallest trees in the segments calculated from the local maxima of the canopy height model. Secondly, this segmentation approach is combined with a special stem detection method. Stem positions in the segments of the watershed segmentation are detected by hierarchically clustering points below the crown base height and reconstructing the stems with a robust RANSAC-based estimation of the stem points. Finally, a new three-dimensional (3D) segmentation of single trees is implemented using normalized cut segmentation. This tackles the problem of segmenting small trees below the canopy height model. The key idea is to subdivide the tree area in a voxel space and to set up a bipartite graph which is formed by the voxels and similarity measures between the voxels. Normalized cut segmentation divides the graph hierarchically into segments which have a minimum similarity with each other and whose members (= voxels) have a maximum similarity. The solution is found by solving a corresponding generalized eigenvalue problem and an appropriate binarization of the solution vector. Experiments were conducted in the Bavarian Forest National Park with conventional first/last pulse data and full waveform LIDAR data. The first/last pulse data were collected in a flight with the Falcon II system from TopoSys in a leaf-on situation at a point density of 10 points/m². Full waveform data were captured with the Riegl LMS-Q560 scanner at a point density of 25 points/m² (leaf-off and leaf-on) and at a point density of 10 points/m² (leaf-on). The study results prove that the new 3D segmentation approach is capable of detecting small trees in the lower forest layer. So far, this has been practically impossible if tree segmentation techniques based on the canopy height model were applied to LIDAR data. Compared to a standard watershed segmentation procedure, the combination of the stem detection method and normalized cut segmentation leads to the best segmentation results and is superior in the best case by 12%. Moreover, the experiments show clearly that using full waveform data is superior to using first/last pulse data.