SVM用于LiDAR数据的地物分类

LiDAR具有探测地表垂向结构的能力,目前还无法被其他遥感源所取代。本文提出通过变换点云提取LiDAR纹理特征,利用支持向量机(SVM)进行训练和分类,并与基于神经网络的分类方法进行比较。试验结果表明,SVM能在训练精度和推广能力之间取得折中,可有效地避免LiDAR地物分类证据不完备引起的过拟合问题,适合于LiDAR地物分类。     

基于层次分析和神经网络的机载LiDAR点云分类

在机载LiDAR点云数据处理中,由于机载雷达点云数据的离散性、不确定性等原因,导致点云分类上很难得到准确的结果。针对机载LiDAR点云数据分类问题,提出了基于层次分析和神经网络的机载LiDAR点云分类方法。根据机载LiDAR点云的数据特征以及不同地物的属性,采用层次分析法赋予每个点云一个二进制信号,然后采用后向传播神经网络(BP-ANN)对机载LiDAR点云数据分类。实验表明:这种方法能够从机载LiDAR独立数据源中分类出房屋、高大的树、低矮的树、道路等地物点云。     

LiDAR数据和航摄影像结合的地物分类研究

对融合机载LiDAR和航摄影像两种数据源地物分类的方法进行了研究,目标是结合LiDAR数据和影像这两种数据各自的特点,综合利用摄影测量、遥感图像处理的相关理论实现典型地物的智能化处理与分类。     

机载多光谱LiDAR数据的地物分类方法

机载多光谱LiDAR系统能够快速地获取大范围地表面上地物光谱和几何数据,并能够保证所获取的光谱与空间几何数据在空间和时间上相对完整和一致性。支持向量机（SVM）是一种基于小样本的学习方法,它避开了从归纳到演绎的传统分类过程。因此,本文提出了基于SVM多光谱LiDAR数据的地物目标分类方法。该方法首先将多个独立波段的LiDAR数据融合为单一的、包含多个波段信息的点云数据,然后将融合后的点云内插为距离影像和多光谱影像,最后利用SVM进行多光谱LiDAR数据的地物覆盖分类。通过对加拿大Optech公司的Titan机载多光谱LiDAR数据的试验证明：相对于传统的单波段LiDAR数据,多光谱LiDAR数据可以获得较好的地物分类精度;比较试验发现SVM分类方法适用于多光谱LiDAR数据的地物分类。     

基于车载LiDAR点云的地物分类方法的研究

基于点云分类常用的近邻聚类法和物体表面分割等方法,本文提出了一种基于最大网格密度的近邻聚类的方法。该方法首先对原始点云进行低点提取,设置格网的大小,在此基础上对点云数据进行去噪并进行主成分分析,再对点云空间进行均匀格网化,使具有最大密度的格网为聚类中心,加入高程、强度以及法向量等特征对分割后的点云实现了不同地物的分类,提高了运算效率,降低了错分率。     

机载多光谱LiDAR的随机森林地物分类

机载多光谱LiDAR技术利用激光进行探测和测距,不仅可以快速获取地面物体的三维坐标,还可以获得多个波段的地物光谱信息,可广泛用于地形测绘、土地覆盖分类、环境建模、森林资源调查等。本文提出了多光谱LiDAR的随机森林地物分类方法。该方法通过对LiDAR强度数据和高程数据提取分类特征,完成多光谱LiDAR的随机森林地物分类;并分析随机森林的特征贡献度特性,采用后向特征选择方法实现分类特征选择。通过对加拿大Optech Titan多光谱LiDAR数据的试验表明：随机森林方法可以获得较好的地物分类精度,而且可以适当地去除部分冗余和相关的特征,从而有效提高分类精度。     

车载LiDAR点云数据中杆状地物自动提取与分类

针对城市道路场景中车载LiDAR点云数据质量差、各类地物相互遮掩的情况,提出杆状地物自动提取与分类算法。先通过改进数学形态学算法移除点云数据中的地面点,再根据杆状地物的形态特征,使用纵向格网模板初步提取杆状地物,然后对提取的疑似杆状地物进行点云数据规则化并通过统计分析移除噪声点,最后根据预先建立的杆状地物样本训练SVM分类模型,对提取的杆状地物进行分类。试验表明,本文方法能够在数据质量欠佳的情况下有效提取城市道路场景中的杆状地物,并对提取的杆状地物进行高精度分类。     

融合航空影像的LiDAR地物点云分类

点云滤波分类是LiDAR后续应用的基础工作,在点云滤波的基础上,以航空影像为辅助条件,结合点云高程信息,设计一套地物点云的分类方法。该方法首先融合航空影像与LiDAR数据,将对应RGB值赋予每个点,根据植被的光谱特征提取出部分植被点云;然后再根据文中定义的点云高程纹理,在剩余地物点云中提取出建筑物点,最后根据回波次数信息分离出剩余植被点,完成地物点云的分类。采用北京凤凰岭地区一组机载LiDAR数据进行实验。实验结果表明,该方法能够有效地将地物点云进行分类并且满足一定的精度要求,具有一定的实用价值。     

基于遗传算法的影像纹理分类

阐述了遗传算法的基本原理,提出与马尔柯夫随机场相结合的遗传算法在影像纹理分类中的应用。通过对４种纹理取样的实验,表明这种影像纹理分类方法可以提高分类的正确率。     

关于BP神经网络转换GPS高程的若干问题

BP神经网络用于GPS高程转换有着较高的精度,但也存在不少问题,如网络的隐含层和隐含层单元个数选取、参加学习的样本的质量如何影响仿真精度等。结合实例分析上述问题,得出一些结论。     

Tree species biomass and carbon stock measurement using ground based-LiDAR

This study scrutinises the use of terrestrial laser scanning (TLS) to measure diameter at breast height (DBH) and tree height at individual tree species level. LiDAR point cloud scans are collected from uniformly defined control points. The result of processed TLS data demonstrates the precise measurements of tree height and DBH by comparing it with field data (DBH, tree height, tree species and location). The average tree height and DBH obtained through TLS measurements were 9.44?m and 43.30?cm, respectively. A linear equation between TLS derived parameters and field measured values were established, which gave the coefficient of determination (r2) of 0.79 and 0.96 for tree height and DBH, respectively. Further, these parameters were used to calculate above ground biomass (AGB) for individual tree species by considering a non-destructive approach. The total AGB and carbon stock from 80 different trees are computed to be 49.601 and 22.320?tonnes, respectively.     

Retrieval of effective leaf area index (LAIe) and leaf area density (LAD) profile at individual tree level using high density multi-return airborne LiDAR

As an important canopy structure indicator, leaf area index (LAI) proved to be of considerable implications for forest ecosystem and ecological studies, and efficient techniques for accurate LAI acquisitions have long been highlighted. Airborne light detection and ranging (LiDAR), often termed as airborne laser scanning (ALS), once was extensively investigated for this task but showed limited performance due to its low sampling density. Now, ALS systems exhibit more competing capacities such as high density and multi-return sampling, and hence, people began to ask the questions like—“can ALS now work better on the task of LAI prediction?” As a re-examination, this study investigated the feasibility of LAI retrievals at the individual tree level based on high density and multi-return ALS, by directly considering the vertical distributions of laser points lying within each tree crown instead of by proposing feature variables such as quantiles involving laser point distribution modes at the plot level. The examination was operated in the case of four tree species (i.e. Picea abies, Pinus sylvestris, Populus tremula and Quercus robur) in a mixed forest, with their LAI-related reference data collected by using static terrestrial laser scanning (TLS). In light of the differences between ALS- and TLS-based LAI characterizations, the methods of voxelization of 3D scattered laser points, effective LAI (LAIe) that does not distinguish branches from canopies and unified cumulative LAI (ucLAI) that is often used to characterize the vertical profiles of crown leaf area densities (LADs) was used; then, the relationships between the ALS- and TLS-derived LAIes were determined, and so did ucLAIs. Tests indicated that the tree-level LAIes for the four tree species can be estimated based on the used airborne LiDAR (R² = 0.07, 0.26, 0.43 and 0.21, respectively) and their ucLAIs can also be derived. Overall, this study has validated the usage of the contemporary high density multi-return airborne LiDARs for LAIe and LAD profile retrievals at the individual tree level, and the contribution are of high potential for advancing forest ecosystem modeling and ecological understanding.     

Estimation of hourly and daily evapotranspiration and soil moisture using downscaled LST over various urban surfaces

Surface moisture is important to link land surface temperature (LST) to people’s thermal comfort. In urban areas, the surface roughness from buildings and urban trees impacts wind speed, and consequently surface moisture. To find the role of surface roughness in surface moisture estimation, we developed methods to estimate daily and hourly evapotranspiration (ET) and soil moisture, based on a case study of Indianapolis, Indiana, USA. In order to capture the spatial and temporal variations of LST, hourly and daily LST was produced by downscaling techniques. Given the heterogeneity in urban areas, fractions of vegetation, soil, and impervious surfaces were calculated. To describe the urban morphology, surface roughness parameters were calculated from digital elevation model (DEM), digital surface model (DSM), and Terrestrial Light Detection and Ranging (LiDAR). Two source energy balance (TSEB) model was employed to generate ET, and the temperature vegetation index (TVX) method was used to calculate soil moisture. Stable hourly soil moisture fluctuated from 15% to 20%, and daily soil moisture increased due to precipitation and decreased due to seasonal temperature change. ET over soil, vegetation, and impervious surface in the urban areas yielded different patterns in response to precipitation. The surface roughness from high-rise has bigger influence on ET in central urban areas.     

Modelling high arctic percent vegetation cover using field digital images and high resolution satellite data

In this study, digital images collected at a study site in the Canadian High Arctic were processed and classified to examine the spatial-temporal patterns of percent vegetation cover (PVC). To obtain the PVC of different plant functional groups (i.e., forbs, graminoids/sedges and mosses), field near infrared-green-blue (NGB) digital images were classified using an object-based image analysis (OBIA) approach. The PVC analyses comparing different vegetation types confirmed: (i) the polar semi-desert exhibited the lowest PVC with a large proportion of bare soil/rock cover; (ii) the mesic tundra cover consisted of approximately 60% mosses; and (iii) the wet sedge consisted almost exclusively of graminoids and sedges. As expected, the PVC and green normalized difference vegetation index (GNDVI; (R_NIR − R_Green)/(R_NIR + R_Green)), derived from field NGB digital images, increased during the summer growing season for each vegetation type: i.e., ∼5% (0.01) for polar semi-desert; ∼10% (0.04) for mesic tundra; and ∼12% (0.03) for wet sedge respectively. PVC derived from field images was found to be strongly correlated with WorldView-2 derived normalized difference spectral indices (NDSI; (R_x − R_y)/(R_x + R_y)), where R_x is the reflectance of the red edge (724.1 nm) or near infrared (832.9 nm and 949.3 nm) bands; R_y is the reflectance of the yellow (607.7 nm) or red (658.8 nm) bands with R²’s ranging from 0.74 to 0.81. NDSIs that incorporated the yellow band (607.7 nm) performed slightly better than the NDSIs without, indicating that this band may be more useful for investigating Arctic vegetation that often includes large proportions of senescent vegetation throughout the growing season.     

结合深度学习的高光谱与多源遥感数据融合分类

高光谱数据具有丰富的光谱特征,但是其空间分辨率相对较低.一些遥感数据具有与高光谱数据互补的优势,例如提供更精细的空间信息的高空间分辨率数据和具有高度信息的激光雷达LiDAR(Light Detection and Ranging)数据.通过将高光谱数据与多源遥感数据进行融合,可以弥补高光谱数据空间分辨率相对较低,空间特...