多时相HJ-1数据补充的水体分布时序变化监测

由于地域、天气等原因造成的云影问题,导致了水体时序分析困难。本文以资源三号（ZY-3）影像作为辅助数据,以环境与灾害监测预报小卫星影像（HJ-1）作为研究数据,通过提取一个时间段内多景时相相近HJ-1数据中有效水体信息,迭代补充成完整的水体提取结果,将不同时间段内影像水体提取的完整结果形成时序监测数据,以达到对某地区水体时序变化监测的目的。通过对淮河流域安徽段研究区水体分布监测表明,该方法能充分利用每个时间段内质量不理想的源数据,对相近时相数据迭代补充,完整提取该时间段内的水体信息。多个时间段迭代补充形成的时序数据可有效支持时序监测。本文在研究区共选取了8295个查找点,结果表明：2013年该研究区洪季（7、8月）水体信息比旱季（3、4月）更丰富,特别是东南部在洪季形成了许多零散水体;相较于旱季,洪季研究区水域面积增长了22.1%。     

青藏高原典型植被生长季遥感模型提取分析

物候变化是衡量全球气候变化最直接、敏感的指示器,针对青藏高原这个独特地域单元上特殊的高寒植被进行关键物候期遥感提取模型及植被物候时空变化的研究具有重要的意义。本文首先以反距离加权空间插值算法与Savitzky-Golay滤波算法相结合的数据重建模型获得高质量2003-2012年青藏高原MODIS归一化植被指数(NDVI)数据。在此数据基础上,分别利用动态阈值法、最大变化斜率法、logistic曲线拟合法3种遥感植被生长季提取模型,对青藏高原地区两种典型植被的生长季(SOS生长季开始期,EOS生长季结束期,LOS生长季长度)进行提取。通过对3种模型提取结果的对比分析,并结合日均温模型对提取结果的验证发现,动态阈值法为青藏高原地区典型植被生长季的最优遥感提取模型。该模型对近10 a的高分辨率典型高寒植被物候参量的反演及时空变化特征分析表明,受青藏高原水热及海拔梯度的影响,青藏高原植被物候变化呈现出从东南向西北的空间分异规律,随春季温度的升高,近10 a来青藏高原高寒草地总体呈现生长季开始期(SOS)提前(0.248 d/a)的趋势。     

融合GNSS气象参数的BP神经网络雾霾预测研究

结合IGS中心获取的BJFS站气象参数（气温（T）、气压（P）、大气可降水量（PWV））及同期PM_2.5数据,建立一种融合时序网络和回归网络的雾霾预测模型,对PM_2.5浓度进行预测。研究表明,引入GNSS气象参数的融合网络模型较单一网络模型适应性强、准确度高,在一定精度范围内可准确预测PM_2.5的变化,时效性达3 h。本文结论验证了卫星导航技术应用于雾霾天气监测及预报的可行性。     

LSTM支持下时序Sentinel-1A数据的太白山区植被制图

植被分类是森林资源调查与动态监测的基础与前提。当前植被分类研究大都利用光学遥感影像,然而,光学遥感成像易受到云雨覆盖的影响,难以构建完整时间序列,植被分类精度有限。微波遥感具有全天时全天候、时间序列完整的优势,在植被调查与分析中具有巨大的应用潜力。本文利用2018年Sentinel-1A微波遥感时间序列数据和深度循环网络方法,对秦岭太白山区的森林植被进行分类制图。首先利用Sentinel-2光学影像与数字高程数据对研究区进行多尺度分割;然后将处理后的时间序列Sentinel-1A数据空间叠加到分割地块上,构建地块的多元时间序列曲线;最后利用深度循环网络提取与学习多元时间序列的时序特征并分类。实验结果表明：① 与传统机器学习方法（如RF、SVM）相比,本文提出的深度循环网络方法的分类精度提高10%以上;② 在Sentinel-1A微波极化特征组合中VV+VH表现最好,与VV+VH+VV/VH极化特征组合的精度相近;③ 使用全年的时间影像构建时间序列分类精度最高,达到82%。研究表明,利用深度循环网络与时间序列Sentinel-1A数据的方法能够有效提高植被分类的精度,从数据源与分类方法上为森林植被分类研究提供了新的思路。     

高分辨率影像城市植被自动提取算法

近十几年来,随着城市化进程加剧,准确获取城市植被的分布信息,是城市气候和地表能量平衡研究的重要内容。高空间分辨率遥感影像数据,为精确获取和动态监测城市植被提供了重要资料。本研究利用资源三号数据对长江三角洲地区城市植被进行光谱特征分析与提取,提出一种城市植被的自动化信息提取算法—分离面法（Hyperplanes for Plant Extraction Methodology,HPEM）。结果表明：在假彩色反射率空间,植被与NDVI值低的背景有很好的分离性,而在真彩色反射率空间,植被与NDVI值高的背景有很好的分离性;HPEM能很好地避免NDVI最佳阈值法中将建筑物误分为植被的问题,其精度明显优于NDVI最佳阈值法,Kappa系数从0.85提高到0.90,总的错分与漏分误差从21.15%降低到14.18%。可见,本文的HPEM方法能有效提高城市植被信息自动提取的精度。     

基于GF-3和Landsat 8遥感数据的土壤水分反演研究

基于我国首颗全极化雷达卫星高分三号（GF-3）和Landsat8数据,研究浓密植被覆盖地表土壤水分反演方法。为了提高浓密植被覆盖地表土壤水分反演精度,首先利用PROSAIL模型、实测植被参数及Landsat8光学数据分析了8种植被指数与植被冠层含水量的相关性,从中优选出归一化差异水指数（NDWI₅）用于反演植被冠层含水量,并通过分析植被含水量和植被冠层含水量的关系,构建植被含水量模型;然后结合植被含水量反演模型和简化MIMICS模型校正了植被对雷达后向散射系数的影响,最后基于AIEM建立裸土后向散射系数模拟数据集,发展一种主动微波和光学数据协同反演浓密植被覆盖地表土壤水分模型,并以山东省禹城市为研究区,实现了玉米覆盖下HH、VV和HH+VV 3种模式土壤水分反演。实验结果表明： ① NDWI₅为最佳植被指数,对于去除植被影响有较好效果;② 基于此方法,利用GF-3和Landsat8卫星数据反演得到的土壤水分具有较高的精度;③ 相比HH和VV两种极化模式,HH+VV双通道模式对土壤水分反演结果更好,决定系数（R2）为0.4037,均方根误差（RMSE）为0.0667 m ³m ^-3。     

水土流失区马尾松林植被提取的土壤调节指数分析

为实现水土流失区植被遥感信息的准确提取,本文采用2007年ALOS 10 m多光谱影像,利用土壤调节植被指数SAVI和MSAVI,对福建长汀水土流失区马尾松林不同植被覆盖密度的3个实验区进行植被提取,并选用不同的土壤调节因子（L=0.25,0.5,0.75,1）做实验,将结果和以NDVI植被指数提取的结果进行对比,分析了提取效果及受土壤噪音的影响程度。实验表明,SAVI指数能提高水土流失区的植被提取精度。在中、低植被覆盖区,其提取的总精度比NDVI高出2%~7%,Kappa系数高出7%~18%;而土壤调节因子L的取值对植被信息的提取也呈现出一定的规律性,即：随着L从0向1递增,SAVI提取稀疏植被的能力上升而探测阴坡植被的能力下降。总体来看,对于低植被覆盖和中等植被覆盖地区,可分别用SAVI（L取0.75）和SAVI（L取0.5）来提取植被信息,对于高植被覆盖区,仍可直接用NDVI进行植被信息提取;研究发现MSAVI在植被信息提取中并不具有特别的优势。     

藏北地区三种时序NDVI重建方法与应用分析

遥感植被指数时间序列数据集,已广泛应用于陆地生态环境变化监测与模拟、植被覆盖动态变化分析、植被物候特征识别与信息提取等多方面的研究.但其因受遥感器采集与传输过程、大气条件、地面状况等诸多因素的影响,时序NDVI数据包含各种噪声,因此研究者们发展了一系列时间序列曲线重建方法.本文对近年来提出或改进的重建算法原理、优缺点进...     

MODIS EVI时序数据重建方法及拟合分析

植被遥感监测中长时间序列数据择优的重建方法,已成为当今一个研究热点。本文以东北地区5种主要植被覆盖类型为例,在定性分析TIMESAT提供的3种常用重建方法对EVI（Enhanced Vegetation Index）时序曲线重建效果的基础上,定量对比研究了各方法,对原始高质量EVI点真实值的保真性,及对原始曲线整体特征的保持度。结果表明：S-G（Savitzky-Golay）滤波对原始曲线生长季的峰值及宽度重建效果较好,但容易因过度拟合保留过多噪声,特别是草地和灌丛类型;非对称性高斯函数（AG）和双Logistic曲线（DL）方法相似,对草地、灌丛和耕地的重建结果更接近真实值,但AG拟合对波峰处异常值的处理结果较差,重建后波峰表现低平。3种算法对原始EVI时序数据的保真性和对原始时序数据曲线特征的保持度,都表现出与植被类型分布相关的空间分布格局。分析结果表明,在东北地区,AG算法对草原和灌丛的重建效果最好,DL算法对耕地重建效果最优,S-G算法最适合对落叶阔叶林和落叶针叶林进行重建处理。     

基于NDVI阈值法的森林冰冻受灾范围精确提取

快速提取森林冰雪受灾范围,有利于准确掌握森林受灾情况,为此类灾害性气候事件防灾减灾、森林资源管理和生态保护提供科学依据。本文利用2001-2007年NDVI数据,提取灾前植被NDVI参考值和正常波动范围,结合2008年NDVI数据提取冰雪冻灾范围。该方法弥补了基于单一时相的传统方法（NDVI差值法）忽略植被指数正常波动的问题,分像元提取植被NDVI正常波动范围,使提取结果更加客观合理。与传统方法提取结果对比,省级尺度的验证结果相同（即森林受灾率均为34.72%,而实地调查森林受灾率35.3%）,但2种方法在县市行政单元提取的森林受灾率相差较大。NDVI阈值法提取的森林冰雪受灾范围主要分布于湖南省南部地区,北部地区分布相对较少,而传统方法提取结果主要分布于湖南省北部地区,南部地区分布相对较少。根据实地考察资料显示,相比于传统方法,NDVI阈值法提取结果与实际森林冰雪冻灾空间分布信息更接近,精度更高,更适合于区域大尺度提取森林冰雪受灾范围。     

The suitability of using leaf area index to quantify soil loss under vegetation cover

Soil erosion by water under forest cover is a serious problem in southern China. A comparative study was carried out on the use of leaf area index (LAI) and vegetation fractional coverage (VFC) in quantifying soil loss under vegetation cover. Five types of vegetation with varied LAI and VFC under field conditions were exposed to two rainfall rates (40 mm h⁻¹ and 54 mm h⁻¹) using a portable rainfall simulator. Runoff rate, sediment concentration and soil loss rate were measured at relatively runoff stable state. Significant negative exponential relationship (p < 0.05, R² = 0.83) and linear relationship (p < 0.05, R² = 0.84) were obtained between LAI and sediment concentration, while no significant relationship existed between VFC and sediment concentration. The mechanism by which vegetation canopy prevents soil loss was by reducing rainfall kinetic energy and sediment concentration. LAI could better quantify such a role than VFC. However, neither LAI nor VFC could explain runoff rate or soil loss rate. Caution must be taken when using LAI to quantify the role of certain vegetation in soil and water conservation.     

Landsat时序变化检测综述

时序变化检测已成为当前Landsat数据主流的变化检测方法。本文从检测算法对比、时序数据构建和精度评价等方面对Landsat时序变化检测进行回顾和评述,进而提出Landsat时序变化检测当前所存在的问题,及其所面临的挑战。Landsat时序变化检测算法可大致归纳为轨迹拟合法、光谱-时间轨迹法、基于模型的方法3大类,这些算法大多基于森林扰动提出;变化检测常用指标有波段型、植被指数型、线性变换型、组合型4大类,每类指标的优势不同,可综合多类指标以更全面地检测不同扰动类型。尽管Landsat时序变化检测已取得长足发展,但仍然面临诸多挑战,其中最大挑战是缺少一致性的参考数据集进行变化检测精度评价。     

Regional-scale identification of three-dimensional pattern of vegetation landscapes

The altitudinal pattern of vegetation is usually identified by field surveys,however,these can only provide discrete data on a local mountain.Few studies identifying and analyzing the altitudinal vegetation pattern on a regional scale are available.This study selected central Inner Mongolia as the study area,presented a method for extracting vegetation patterns in altitudinal and horizontal directions.The data included a vegetation map at a 1∶1 000 000 scale and a digital elevation model at a 1∶250 000 scale.The three-dimensional vegetation pattern indicated the distribution probability for each vegetation type and the transition zones between different vegetation landscapes.From low to high elevations,there were five vegetation types in the southern mountain flanks,including the montane steppe,broad-leaved forest,coniferous mixed forest,montane dwarf-scrub and sub-alpine shrub-meadow.Correspondingly,only four vegetation types were found in the northern flanks,except for the montane steppe.This study could provide a general model for understanding the complexity and diversity of mountain environment and landscape.     

森林过火区植被遥感参数的变化与恢复特征分析

遥感技术可以快速、准确地监测森林火灾火烧迹地的植被遥感参数变化,分析植被对火灾的响应与恢复特征,为防灾减灾决策提供科学依据。本文首先基于森林火灾前后的Landsat5 TM数据,利用差分归一化燃烧指数（the Differential Normalized Burn Ratio,dNBR）来提取2009年澳大利亚维多利亚州火烧迹地的范围,计算过火区面积及火烧强度;其次基于时间序列的全球地表特征参量（Global Land Surface Satellite,GLASS）产品中的叶面积指数（Leaf Area Index,LAI）、吸收光合有效辐射比例（Fraction of Absorbed Photosynthetically Active Radiation,FAPAR）数据,利用距平分析法对比不同火烧强度过火区植被与未过火区植被受森林火灾的影响状况与植被恢复特征。结果表明,森林火灾发生后,LAI、FAPAR值迅速降低,火烧强度越大,LAI、FAPAR下降程度越大,高火烧强度过火区的LAI、FAPAR最大降幅分别为中火烧强度、低火烧强度过火区的1.2、1.3倍;随时间推移,LAI、FAPAR值逐渐上升,在2-3年内恢复至未过火区水平。LAI、FAPAR恢复至未过火区平均水平的时间与森林火灾规模、火烧强度密切相关：维多利亚州森林火灾过火区域中大过火斑块、高火烧强度林地的植被遥感参数恢复时间相比小过火斑块、低火烧强度林地滞后1-2年。植被遥感参数LAI、FAPAR能很好地反映过火区植被的受损状况及恢复过程。     

基于SSA-Mann Kendall的草原露天矿区NDVI时间序列分析

针对煤矿区植被指数时间序列研究中,存在年际尺度上对植被动态规律刻画不全面、月际尺度上因物候性周期波动导致变化趋势和周期振荡信号微弱难以提取、基于变换的变化检测物理意义不够明确的问题,本文以胜利露天矿区为例,在月际尺度,基于SSA-Mann Kendall重建草原露天矿区的采矿扰动区和伪不变特征区MODIS NDVI时间序列的趋势和周期振荡信号,从显著程度和突变时间2方面对趋势成分进行定量化分析,并结合各特征区周期振荡演变特征揭示采矿扰动下草原露天矿区植被生长的动态规律。结果表明：SSA-Mann Kendall能将NDVI时间序列的微弱信号充分放大,便于提取,并可对趋势成分进行定量化表达,结合周期振荡与趋势成分的演变特征有助于辅助识别矿区植被生长的动态特点;伪不变特征区植被无显著下降趋势,采矿扰动区下降趋势显著,且露天采场较排土场的趋势更为明显,草原露天矿区地表植被损伤具有突发性,突变点多发于矿井开工建设时;扰动形式差异导致部分矿井露天采场和排土场周期振荡演变特征存在差异,露天采场植被消失殆尽,排土场因植被恢复措施而具有更复杂的动态特点。     

随机森林算法在机载LiDAR数据林分平均树高 估算中的应用研究

采用机载LiDAR数据估算森林结构参数是当前林业遥感中的研究热点。本文以福建省长汀县朱溪河流域为示范区,探讨了随机森林算法（RF）在机载LiDAR数据林分平均树高估测中的适用性。首先,通过渐进三角网（TIN）算法进行点云滤波并获取相应林分样地的植被点云子集和高程归一化的植被点云;然后,从归一化后的植被点云提取出高度分位数变量以及点云统计特征值等24个变量参数;最后,根据提取的变量参数和野外实测林分均高数据建立研究区林分平均高随机森林回归估测模型。研究结果表明,模型估测的样地平均树高与实测值具有明显线性相关关系,线性回归系数为0.938,相关系数达到0.968。对样地的估测精度都在86%以上,总体平均精度达到了93.17%。研究认为,基于植被点云变量参数的随机森林模型估测林分平均树高具有较高的可靠性。     

Effect of water level fluctuations on wintering goose abundance in Poyang Lake wetlands of China

The Poyang Lake is a Ramsar site and is the important over-wintering site for migratory waterbirds along the East Asian–Australasian Fly way. Examining the effects of water level fluctuations on waterbird abundance and analyzing the influencing mechanism is critical to waterbird protection in the context of hydrological alteration. In this study, the effect of water level regime on wintering goose abundance was examined and the influencing mechanism was interpreted. Synchronous waterbirds survey data, hydrological data, Moderate Resolution Imaging Spectroradiometer-Normalized Difference Vegetation Index(MODIS-NDVI) data and habitat data derived from Landsat TM/ETM data and HJ/CCD data were combined. The satellite-derived Green Wave Index(GWI) based on MODIS-NDVI dataset was applied to detect changes in goose food resources. It was found that habitat size and vegetation conditions are key factors determining goose abundance. Geese numbers were positively correlated with habitat area, while intermediate range of vegetation productivity might benefit the goose abundance. Water level affects goose abundance by changing available habitat areas and vegetation conditions. We suggested that matching hydrological regime and exposed meadows time to wintering geese dynamics was crucial in the Poyang Lake wetlands. Our study could provide sound scientific information for hydrological management in the context of waterbird conservation.     

Using multispectral landsat and sentinel-2 satellite data to investigate vegetation change at Mount St. Helens since the great volcanic eruption in 1980

Long-term analyses of vegetation succession after catastrophic events are of high interest for an improved understanding of succession dynamics. However, in many studies such analyses were restricted to plot-based measurements. Contrarily, spatially continuous observations of succession dynamics over extended areas and time-periods are sparse. Here, we applied a change vector analysis (CVA) to investigate vegetation succession dynamics at Mount St. Helens after the great volcanic eruption in 1980 using Landsat. We additionally applied a supervised random forest classification using Sentinel-2 data to map the currently prevailing vegetation types. Change vector analysis was performed with the normalized difference vegetation index (NDVI) and the urban index (UI) for three subsequent decades after the eruption as well as for the whole observation time between 1984 and 2016. The influence of topography on the current vegetation distribution was examined by comparing altitude, slope angles and aspect values of vegetation classes derived by the random forest classification. Wilcox- Rank-Sum test was applied to test for significant differences between topographic properties of the vegetation classes inside and outside of the areas affected by the eruption. For the full time period, a total area of 516 km² was identified as re-vegetated, whereas the area and magnitude of re-growing vegetation decreased during the three decades and migrated closer to the volcanic crater. Vegetation losses were mainly observed in regions unaffected by the eruption and related mostly to timber harvesting. The vegetation type classification reached a high overall accuracy of approximately 90%. 36 years after the eruption, coniferous and deciduous trees have established at formerly devastated areas dominating with a proportion of 66%, whereas shrubs are more abundant in riparian zones. Sparse vegetation dominates at regions very close to the crater. Elevation was found to have a great influence on the reestablishment and distribution of the vegetation classes within the devastated areas showing in almost all cases significant differences in altitude distribution. Slope was less important for the different classes - only representing significantly higher values for meadows, whereas aspect seems to have no notable influence on the reestablishment of vegetation at Mount St. Helens. We conclude that major vegetation succession dynamics after catastrophic events can be assessed and characterized over large areas from freely available remote sensing data and hence contribute to an improved understanding of succession dynamics.