无人机激光点云和多光谱数据的融合技术研究

该文研究无人机激光点云数据和多光谱通道树种数据的融合和树种智能判别,综合利用多光谱数据与激光雷达数据的数据优势,多景的多光谱数据经过图像配准、平差计算后,精确地将影像数据对准到大地坐标系中,再以三维的激光点云数据为基准,通过数据压缩、多光谱数据写入,实现对多光谱数据与激光雷达数据的有效融合,为输电线路巡检提供良好的数据基础。     

机载LIDAR点云数据估测单株木生物量

使用黑河流域遥感-地面观测同步试验获取的机载激光雷达(LIDAR)点云数据估测了典型树种的单株木生物量。首先由点云数据生成冠层高度模型(CHM),然后采用优化的单株木树冠特征识别算法估测相关的结构参数,最后通过回归分析建立估测参数(树高、冠幅)与实测参数(树高、冠幅、胸径)之间的最优回归方程,并与现有的单株木生物量实测相关生长方程联立,得到单株木生物量估测相关生长方程。结果表明,由LIDAR点云数据得到的单株木估测参数与实测参数显著相关,可以估测单株木生物量(R~2为0.729)。     

基于卷积神经网络的食品塑料包装袋光谱识别

目的实现食品塑料包装袋的快速检测和材质区分。方法研究使用高光谱成像技术在450～950nm波长范围下采集了49组不同食品包装袋样本的光谱数据,利用Savitzky-Golay平滑滤波、数据归一化和主成分分析进行预处理,建立决策树、支持向量机2种传统机器学习模型和卷积神经网络模型,并比较了它们对包装袋材质的识别性能。结果决策树模型与支持向量机模型的验证识别率分别为87.8%和88.9%,卷积神经网络模型的验证识别率高达100%,损失函数值最终下降到0.0171且达到收敛,在分类效果和精度上具有明显的优势。结论高光谱检测方法不破坏检材,重现性好,稳定性强,实现了对食品塑料包装袋的精准识别。卷积神经网络模型对食品包装袋高光谱数据的识别效果最好,为食品包装袋质量检测领域中塑料包装袋的识别鉴定提供依据。     

基于地基激光雷达数据的单木结构参数提取研究

针对利用地基激光雷达数据提取单木结构参数的问题,提出了一套较为复杂林分条件下的地基激光雷达数据获取方案与单木结构参数提取方法。首先利用八邻点距离检验法对点云数据中的多次散射点进行滤除,进而通过对现有的变尺度地面点识别方法进行改进,实现了对研究区内的地面点云的快速识别。通过对地基激光雷达数据的分析,提出了一种基于树干点云垂直连续分布特征的树干识别方法,在此基础上提取了单木位置、胸径和树高,最后将所提取的结构参数与林业测量数据进行了对比,结果具有较好的一致性。     

基于遥感专家分类系统的土壤信息提取与识别

论文在遥感信息提取与识别理论与方法的支持下，以黑龙江省富裕县为试验区，采用MODIS遥感数据和地形数据，利用地表反射率、植被指数以及DEM数据等参数提取特征图像．结合样区土壤分类体系和土壤遥感光谱特征，探讨基于专家分类系统的土壤信息提取的可行性，评价其识别的精度。     

基于激光雷达数据的火星表面障碍物识别

火星车的环境感知能力是其进行智能移动和探测的基础,而障碍物检测识别是环境感知中的一个重要方面,识别效果直接决定了火星车工作能力和安全性。本文提出一种基于激光雷达数据的火星表面障碍物自动识别方法。通过获取的激光雷达点云数据,首先在分析激光反射强度理论的基础上,通过强度补偿理论将点云强度根据距离、角度因素进行修正,进而构建激光雷达强度值与目标特征的反射关系。通过大津法自动求取全局阈值,自适应的将火星表面点云分类为障碍物点云和非障碍物点云;然后通过曲率约束剔除不符合条件的障碍物点云;最后利用基于八叉树叶节点的连通性聚类,实现火星表面障碍物点云的识别。模拟实验结果表明,该方法可实现激光雷达点云中的火星表面障碍物有效提取,典型障碍物识别精度接近90%,为基于火星车障碍物检测和环境感知相关研究提供借鉴。     

基于人工神经网络的光谱反射率重建

目的研究基于BP神经网络法和FNN神经网络法重构图像光谱反射率的精度。方法以SG标准色卡作为训练样本,分别使用BP和FNN神经网络法,对测试样本DC标准色卡的光谱反射率进行预测,并利用CIEL*a*b*色差公式、均方根误差(ERMS)和光谱匹配精度(GFC)对结果进行评价。结果 BP和FNN神经网络重构的光谱反射率平均色差(ΔEab)分别为2.997和3.071,平均均方根误差(ERMS)分别为0.056和0.049,平均光谱匹配精度(GFC)分别为0.987和0.991。结论 2种神经网络方法重构的光谱反射率具有相当优越的色度和光谱精度。相比于FNN神经网络,BP神经网络更加适合于光谱图像的获取领域。     

马尾松良种基地种植技术以及效益分析

马尾松以适应性强和经济价值高的双重营林特性,成为豫南山区荒山造林的先锋树种。但长期以来不合理开发及乱砍滥伐,马尾松资源遭到很大破坏,因此必须建立马尾松良种基地,改良提高种子遗传品质;逐步实现良种生产专业化、采种基地化、造林良种化。     

一种基于宽度学习的高精度光谱重建方法

目的 研究一种更有效的光谱重建方法,以提升光谱重建的精度。方法 文中提出一种基于宽度学习的光谱重建方法,以包含1 269个色块的孟塞尔亚光数据集和包含289个色块的Agfa IT8.2数据集为实验样本,利用商用彩色数码相机的模拟系统对所提方法进行验证,以光谱均方根误差、光谱拟合优度系数和2种色差公式为算法评价指标,并与现有的光谱重建方法进行了对比。结果 实验结果表明,该方法可实现的平均均方根误差低至0.4%,平均光谱拟合优度系数达到99.9%,平均色差低至0.147和0.112,光谱精度和色度精度都明显优于其他2种方法。结论 基于宽度学习的光谱重建算法可以有效地提高光谱重建的精度,能够实现更高精度的光谱颜色表征和再现的要求。     

应用变分模态分解和随机森林特征选择算法的扬声器异常声分类

为了提高扬声器异常声分类的精度,提出了基于变分模态分解(variational mode decomposition, VMD)和随机森林特征选择算法的扬声器异常声分类方法。首先利用VMD分解采集到的扬声器声响应信号,之后对得到的一系列模态分量提取时域和频域特征;然后利用随机森林特征计算提取特征的重要性,通过递归特征消除算法提取出相关性较强的特征构造出最优特征子集;最后将最优特征子集输入至随机森林分类器中,实现扬声器异常声的分类识别。试验结果表明,该方法可以筛选出规模较小且识别度较高的低维特征数据集,同时具有更好的平均识别准确率,平均识别准确率为98.61%。     

Tangent height registration method for the Version 1.4 data retrieval algorithm of the solar occultation sensor ILAS-II

The Improved Limb Atmospheric Spectrometer-II (ILAS-II) is a satellite-borne solar occultation sensor onboard the Advanced Earth Observing Satellite-II (ADEOS-II). The ILAS-II succeeded the ILAS. The ILAS-II used four grating spectrometers to observe vertical profiles of gas volume mixing ratios of trace constituents and was also equipped with a Sun-edge sensor to determine tangent heights geometrically with high precision. The accuracy of gas volume mixing ratios depends on the accuracy of the tangent height determination. The combination method is a tangent height registration method that was developed to give appropriate tangent heights for the ILAS-II Version 1.4 data retrieval algorithm. This study describes the method used in the ILAS-II Version 1.4 retrieval algorithm to register tangent heights. The root-sum-square total random error is estimated to be 30 m, and the total systematic error is 180 m at an altitude of 30 km. The influence of the tangent height errors on the vertical profiles of gas volume mixing ratios in ILAS-II Version 1.4 is estimated by using the relative difference. The relative difference for each species is within 7% (20%) for an altitude shift of +/-100 m(+/-300 m).     

Entrainment of Heterogeneous Particles from Gas-Fluidized Bed

The entrainment of heterogeneous particles in a gas-fluidized bed for particle mixtures of categories A and B according to the Geldart classification was conducted in this study. The experiments were carried out in an acrylic column of 0.092 m diameter. The distributors were perforated plates of up to 5.9% free area fraction. Measurements of pressure gradients were made using 24 pressure taps. Glass beads with particle diameters from 60 to 400 μm were used. The mixtures were described using the Rosin-Rammler-Bennet granulometric distribution model. The transport disengaging height (TDH) heights were obtained using the Geldart methodology. A 2⁵ experimental design was applied relating the dispersion index, mean diameter, solid mass, superficial gas velocity, and free area fraction of the distributor to obtain the TDH heights. Slugging and transition for turbulence regimes characterized the behaviors of fluidized beds. The results showed that TDH heights for heterogeneous particles were dependent on the solid mass and superficial gas velocity.     

LEDs (λ<Subscript>max</Subscript> = 3.6 μm) with cone-shaped light-emitting surfaces

A series of light-emitting diodes (LEDs) operating at λ_max = 3.6 μm are created using cone-shaped mesas with heights of 10–130 μm and concave side surfaces. The dependence of the efficiency of room-temperature (T = 298 K) emission on the mesa height at various injection currents has been studied. The character of the observed dependence agrees with the results of theoretical calculations. The output radiation power of LEDs with the maximum mesa height (130 μm) at a pumping current of 220 mA amounts to 53 μW, which is 1.5 times higher than the power of LEDs with a mesa height of 10 μm.     

发泡聚乙烯最大加速度-静应力曲线快速获取方法研究

目的 研究快速、准确预测最大加速度-静应力曲线的方法。方法 首先利用落锤冲击试验机获取了5个不同高度下,5种不同厚度的发泡聚乙烯的最大加速度-静应力曲线。在此基础上,分析对比文中3种不同的改进拟合法与已有的动应力与动能量多项式拟合法的区别。结果 研究发现,当不区分高度的情况下,以最大加速度因子为函数值,以跌落高度、衬垫厚度、静应力为变量进行拟合时,其代表预测精度R²的平均为0.835,相比动应力与动能量多项式拟合法的0.299 6要高。但曲线右侧的预测精度偏低。引入以静应力为变量的多项式作为修正因子后,R²的平均值为0.934。预测精度有所提高,右侧的预测偏差减小,但仍存在。在区分高度的情况下,以带有修正因子的公式进行预测时,R²的平均值为0.984,曲线向右侧预测偏差逐渐增大的现象明显改善。结论 区分高度情况下,利用带修正因子的预测公式可以快速且较准确地预测最大加速度-静应力曲线,可以为冲击防护设计及相关软件的开发提供一定的帮助。     

基于BP神经网络的导线脱冰跳跃高度预测模型

冰区导线脱冰振动会引起绝缘间隙减小,严重时甚至导致闪络和跳闸等电气事故.首先利用数值方法模拟得到各种参数条件下导线的脱冰动力响应,获得导线的最大脱冰跳跃高度.进而基于数值模拟结果和BP神经网络构建导线脱冰跳跃高度预测模型,将线路的导线分裂数、导线型号、档距、高差等结构参数以及初始应力、覆冰厚度和脱冰率等载荷参数作为输入...     

Is x‐height a better indicator of legibility than type size for drug labels?

In 1999 the US Food and Drug Administration published a regulation in an attempt to ensure the legibility of OTC drugs, specifying, among other things, a minimum type size of 6 points. This is problematic because different typefaces of the same size vary widely in type heights and, presumably, legibility. We hypothesized that specifying a minimum x‐height, the height of the lowercase x, would produce more consistent legibility than the minimum type size specified within the regulation. Twenty‐six subjects viewed two groups of typefaces using the Lockhart Legibility Instrument to quantify legibility. The first group contained typefaces that were all 6 points, but, by nature of their design, varied greatly in their x‐heights. The second group was made from the same set of typefaces, but these were manipulated so that their x‐heights were equal to the average x‐height of group 1. A likelihood ratio test indicated that the group that varied in x‐height, group 1, produced significantly more variable results than the group with equal x‐heights, group 2. This indicates that specifying a minimum type size may not be the best approach for producing consistent legibility. Copyright © 2003 John Wiley & Sons, Ltd.     

Determination of organic and inorganic carbon in forest soil samples by mid-infrared spectroscopy and partial least squares regression

Analyses of organic and inorganic carbon are of great interest in the field of soil analyses. Soil samples from a national monitoring project were provided for this study, including more than 130 forest sites from Austria. We investigated the humus layers (if present undecomposed litter (L), of mixed samples of F- (intermediate decomposed organic matter) and H-(highly decomposed organic matter) (FH)) and upper mineral soil layers (0-5 and 5-10 cm) of the samples. Mid-infrared spectra were recorded and evaluated by their band areas; subsequently we calculated models with the partial least squares approach. This was done by correlating calculated data of the mid-infrared spectra with gas-volumetrically determined carbonate values and measurements of organic carbon from an elemental analyzer. For carbonate determination, this approach gave satisfying results. For measurements of organic carbon, it was necessary to discriminate into humus layers and mineral soils or even more groups to obtain satisfactory correlations between spectroscopically determined and conventionally measured values. These additional factors were the presence of carbonate, the forest type, and the dominant tree species. In mineral soils, fewer subdivisions were necessary to obtain useful results. In humus layers, groupings of sites with more similar characteristics had to be formed in order to obtain satisfying results. The conclusion is that the chemical background of soil organic matter leading to different proportions of functional groups, especially in the less humified organic matter of the humus layers, plays a key role in analyses with mid-infrared spectroscopy. Keeping this in mind, the present approach has a significant potential for the prediction of properties of forest soil layers, such as, e.g., carbonate and organic carbon contents.     

The reason for an upper limit to the height of spinnable carbon nanotube forests

The reason for the upper limit on the height of spinnable carbon nanotube (CNT) forests was studied. To analyze the differences between CNT forests with different heights, we synthesized CNT forests using different growth times (3, 6, 9, 12, 15, and 60 min). The height of the CNT forests increased from 260 μm at 3 min to 1.7 mm at 60 min, and the spinnability decreased sharply after 9 min of growth, where a wavy morphology first appeared. Raman analysis of the CNT forest grown for 9 min showed that the intensity ratio of G-band to D-band at the upper region was 1.50 and that near the bottom was 1.14. We also found that the reaction termination process affected the spinnability of the CNT forests. Depending on the termination process, both spinnable and non-spinnable CNT forests could be selectively synthesized, because of the different morphologies in their lower regions. The results suggested that any wavy morphology produced due to a disturbance in growth conditions causes a loss of spinnability.     

Effects of obstacle height on the control of the body center of mass motion during obstructed gait

Abstract

Tripping over obstacles has been reported as one of the most frequent causes of falls in the elderly. Since a successful crossing requires the stability of the body through the control of the motion of the body center of mass (COM), the purpose of this study was to investigate the COM motion when crossing obstacles of different heights. Ten young healthy adults walked and crossed obstacles of heights of 10%, 20% and 30% of their leg lengths in a gait laboratory. The COM motions were calculated using data measured from a motion analysis system. It was found that vertical motions of the COM, including position, velocity and acceleration, were all modulated to successfully cross obstacles of different height. All acceleration components of the COM were changed with increased obstacle height, suggesting that different forces from muscle contractions were needed to control COM stability when facing different obstacle heights. These findings may help shed light on further investigation of the motor control strategies of the central nervous system for obstacle crossing and serve as baseline data in the evaluation of the elderly and patients.     

Lidar characterization of crystalline silica generation and transport from a sand and gravel plant

Light detection and ranging (Lidar) remote sensing two-dimensional vertical and horizontal scans collected downwind of a sand and gravel plant were used to evaluate the generation and transport of geologic fugitive dust emitted by quarry operations. The lidar data give unsurpassed spatial resolution of the emitted dust, but lack quantitative particulate matter (PM) mass concentration data. Estimates of the airborne PM10 and crystalline silica concentrations were determined using linear relationships between point monitor PM10 and quartz content data with the lidar backscatter signal collected from the point monitor location. Lidar vertical profiles at different distances downwind from the plant were used to quantify the PM10 and quartz horizontal fluxes at 2-m vertical resolution as well as off-site emission factors. Emission factors on the order of 65-110 kg of PM10 (10-30 kg quartz) per daily truck activity or 2-4 kg/t product shipped (0.5-1 kg quartz/t) were quantified for this facility. The lidar results identify numerous elevated plumes at heights >30 m and maximum plume heights of 100 m that cannot be practically sampled by conventional point sampler arrays. The PM10 and quartz mass flux was greatest at 10-25 m height and decreased with distance from the main operation. Measures of facility activity were useful for explaining differences in mass flux and emission rates between days. The study results highlight the capabilities of lidar remote sensing for determining the spatial distribution of fugitive dust emitted by area sources with intermittent and spatially diverse dust generation rates.