激光雷达点云树木建模研究进展与展望

三维树木几何模型是数字城市与数字林业工程的重要组成部分.针对点云树木建模,深入分析了基于广义(泛在)激光雷达点云的树木模型重建方法,提出了聚类思想建模、图论方法建模、先验假设建模、拉普拉斯算子建模与轻量化表达建模5类建模体系,归纳总结了不同建模体系在树冠枝干的细节表达、建模算法性能、树木模型的多层次细节表达、建模体系综...     

Molecular clouds and the formation of open clusters

Using analytical approximations of the mass spectra of molecular clouds and stellar systems including open clusters and associations a relation between the final star formation efficiency (final SFE) and the initial cloud mass is derived. The dependence of the SFE on cloud mass is estimated for two cases of molecular cloud destruction by stellar winds in spherical geometry. The result of growing SFE with rising cloud mass does not agree with current observations and the above relation. A decrease of the SFE is expected if the fraction of high-mass stars grows with increasing mass of the cloud.     

京津冀夏季强降水下冰云宏微观特征

根据Aqua MODIS 2级云产品和Cloudsat的2级产品资料,结合降水数据和MODIS L1B级辐射率数据,对发生在京津冀地区夏季的三次强降水过程中冰云的宏微观物理量的特征进行分析,并探究这些物理量和降水强度的关系。结果表明：在水平分布中,强降水过程中降水强度高值区内云相为冰云,冰云云顶高度在8~17 km,冰云粒子有效半径、冰云光学厚度、冰水路径分别最高可达60 μm、 150、 5 000 g?m^-2;冰云光学厚度、冰水路径、冰云云顶高度随降水强度增大而增大。在垂直分布中,冰云主要分布在3.5 km以上,发生强降水站点的冰云为深对流云,冰云粒子有效半径、冰水含量、冰云粒子数浓度分别最高可达150 μm、 3 000 mg?m^-3 、 500 L^-1;冰云粒子有效半径高值区存在于云层中下部,且随高度上升而减小,冰云粒子数浓度高值区存在于云层中上部,且随高度上升而增加,冰水含量高值区则存在于云层中部;冰云粒子有效半径、冰水含量、冰云粒子数浓度在9 km以上随降水强度增大而增大。     

基于随机森林的光子计数激光雷达点云滤波

新一代星载激光雷达卫星ICESat-2首次采用了微脉冲光子计数激光雷达技术,由于单光子探测的灵敏性导致数据在大气和地表下层产生了大量噪声,因此对光子计数激光雷达点云数据实现信号和噪声的分离是开展进一步应用研究的前提和基础。本文选择美国俄勒冈州和弗吉尼亚州2个研究区,采用MATLAS数据,根据光子点云数据的特点构造了12个光子点云特征,对所构造的特征利用随机森林进行变量筛选,用机器学习方法对光子点云进行分类,并将建立好的模型推广到整个研究区。研究结果表明,本文构建的分类器分类总精度达到了96.79%,Kappa系数为0.94,平均生产者精度和用户精度分别为97.1%和96.8%。在相对弱噪声、平坦地形区域和强噪声、复杂地形区域都取得较好的分类结果。本文结果显示了基于少量样本通过机器学习的方法构建模型,可以推广到较大范围区域的光子点云分类应用中。     

基于加权约束的单体建筑物点云表面重建算法

建筑物点云表面重建在高精度城市测绘、虚拟现实等领域有十分广泛的应用前景。由于建筑物的几何形态多变,重建算法普遍存在计算速率慢、拟合精度低和模型结构不完整的问题。为此,本文以单体建筑物为研究对象,提出基于加权约束的单体建筑物点云表面重建算法,在表面初始化过程中充分考虑数据对结构拟合的贡献。在此基础上,构建基于正则集的单体建筑物表面重建算法,实现建筑物拟合过程中的加权拟合误差、近邻结构平滑的同步优化。针对多类建筑物三维点云的实验结果表明,相比传统的建筑物重建策略,本文的加权约束方法可根据不同类型的点云数据设计自适应权重,并选择模型拟合中最优的权重函数,在高噪声、低精度点云数据下能得到更高精度的单体建筑物表面模型。     

基于点云数据的改进四叉树空间索引研究

随着测绘行业的不断发展,三维激光扫描技术已经成为地理信息产业中不可或缺的重要组成部分^[1],然而点云庞大的数据量给处理带来了一定的困难。本文针对运用车载移动测量系统所获取的海量点云数据,采用了一种基于编码改进四叉树索引的点云数据组织处理方法。结果证明该方法提高了创建索引的速度,减少了树的深度以及数据的冗余量,并基于此编码进行最邻近查找,具有可行性和有效性。     

Radiation pressure — a stabilizing agent of dust clouds in comets?

The internal dynamics of an illuminated dust cloud of finite optical thickness is investigated. The dependence of the radiation pressure on the optical depth makes the individual particles oscillate, in one dimension, around the accelerated centre of gravity of the cloud. The cloud moves as an entity, irrespectively of the velocity dispersion of the particles and their efficiency for radiation pressure. If the optical depth does not change, i.e. if the cloud does not expand laterally, its lifetime is unlimited. A contraction caused by energy dissipation in mechanical collissions between the dust particles is expected. The range of particle sizes which can be transported by such a “coherent cloud” is estimated, as well as the acceleration of the whole cloud. The structure of the cloud in real space and in velocity space is investigated. A comparison with the “striae” observed in the dust tails of great comets shows that the parent clouds of these striae may have been of the kind considered.     

基于“葵花8号”气象卫星的陇东南地区强对流识别跟踪技术研究

利用常规观测资料、区域自动站资料和“葵花8号”气象卫星资料,对2016年4—9月甘肃省陇东南地区出现的43次强对流天气过程进行分析,确立了强对流云团识别指标、追踪方法及预报指标,并对2018年部分个例进行效果检验。结果表明：（1）利用卫星B13通道（10.4μm）亮温值TBB≤238 K或B08通道（6.2μm）与B13通道亮温差△TBB<0 K双阈值作为强对流云团识别指标,可以准确识别出陇东南地区的强对流天气云团;（2）利用“逆向搜索法”、“面积重叠法”及对云团重心的计算,可以对强对流云团进行准确定位、追踪及移动路径外推预报;（3）建立的强对流天气落区判别指标对该地区短时强降水及冰雹落区具有一定的预报能力。     

An assessment of Arctic cloud water paths in atmospheric reanalyses

The role of Arctic clouds in the recent rapid Arctic warming has attracted much attention. However, Arctic cloud water paths(CWPs) from reanalysis datasets have not been well evaluated. This study evaluated the CWPs as well as LWPs(cloud liquid water paths) and IWPs(cloud ice water paths) from five reanalysis datasets(MERRA-2,MERRA, ERA-Interim, JRA-55, and ERA5) against the COSP(Cloud Feedback Model Intercomparison Project Observations Simulator Package) output for MODIS from the MERRA-2 CSP(COSP satellite simulator) collection(defined as M2 Modis in short). Averaged over 1980–2015 and over the Arctic region(north of 60°N), the mean CWPs of these five datasets range from 49.5 g/m~2(MERRA) to 82.7 g/m~2(ERA-Interim), much smaller than that from M2 Modis(140.0 g/m~2). However, the spatial distributions of CWPs, show similar patterns among these reanalyses, with relatively small values over Greenland and large values over the North Atlantic. Consistent with M2 Modis, these reanalyses show larger LWPs than IWPs, except for ERA-Interim. However, MERRA-2 and MERRA underestimate the ratio of IWPs to CWPs over the entire Arctic, while ERA-Interim and JRA-55 overestimate this ratio. ERA5 shows the best performance in terms of the ratio of IWPs to CWPs. All datasets exhibit larger CWPs and LWPs in summer than in winter. For M2 Modis, IWPs hold seasonal variation similar with LWPs over the land but opposite over the ocean. Following the Arctic warming, the trends in LWPs and IWPs during 1980～2015 show that LWPs increase and IWPs decrease across all datasets, although not statistically significant. Correlation analysis suggests that all datasets have similar interannual variability. The study further found that the inclusion of re-evaporation processes increases the humidity in the atmosphere over the land and that a more realistic liquid/ice phase can be obtained by independently treating the liquid and ice water contents.     

The role of clouds in tropospheric photochemistry

We show that photochemical processes in the lower half of the troposphere are strongly affected by the presence of liquid water clouds. Especially CH₂O, an important intermediate of CH₄ (and of other hydrocarbon) oxidation, is subject to enhanced breakdown in the aqueous phase. This reduces the formation of HO _x -radicals via photodissociation of CH₂O in the gas phase. In the droplets, the hydrated form of CH₂O, its oxidation product HCO₂ ^–, and H₂O₂ recycle O₂ ^– radicals which, in turn, react with ozone. We show that the latter reaction is a significant sink for O₃. Further O₃ concentrations are reduced as a result of decreased formation of O₃ during periods with clouds. Additionally, NO _x , which acts as a catalyst in the photochemical formation of O₃, is depleted by clouds during the night via scavenging of N₂O₅. This significantly reduces NO _x -concentrations during subsequent daylight hours, so that less NO _x is available for O₃ production. Clouds thus directly reduce the concentrations of O₃, CH₂O, NO _x , and HO _x . Indirectly, this also affects the budgets of other trace gases, such as H₂O₂, CO, and H₂.