High spatial resolution image object classification for terrestrial oil spill contamination mapping in West Siberia

This work is a part of the OSCaR pilot study (Oil Spill Contamination mapping in Russia). A synergetic concept for an object based and multi temporal mapping and classification system for terrestrial oil spill pollution using a test area in West Siberia is presented. An object oriented image classification system is created to map contaminated soils, vegetation and changes in the oil exploration well infrastructure in high resolution data. Due to the limited spectral resolution of Quickbird data context information and image object structure are used as additional features building a structural object knowledge base for the area. The distance of potentially polluted areas to industrial land use and infrastructure objects is utilized to classify crude oil contaminated surfaces. Additionally the potential of Landsat data for dating of oil spill events using change indicators is tested with multi temporal Landsat data from 1987, 1995 and 2001. OSCaR defined three sub-projects: (1) high resolution mapping of crude oil contaminated surfaces, (2) mapping of industrial infrastructure change, (3) dating of oil spill events using multi temporal Landsat data. Validation of the contamination mapping results has been done with field data from Russian experts provided by the Yugra State University in Khanty-Mansiyskiy. The developed image object structure classification system has shown good results for the severely polluted areas with good overall classification accuracy. However it has also revealed the need for direct mapping of hydrocarbon substances. Oil spill event dating with Landsat data was very much limited by the low spatial resolution of Landsat TM 5 data, small scale character of oil spilled surfaces and limited information about oil spill dates.     

利用无人机高光谱影像的红树林群落物种分类

近年来红树林群落中物种结构简单、功能退化等环境问题日趋严重,为了及时准确掌握红树林群落的物种空间格局与分布,本文首先基于深圳福田红树林自然保护区无人机高光谱影像,利用归一化差值植被指数和归一化潮间红树林指数提取植被区域;然后在植被区域根据最佳指数法选取信息量大、波段相关性小的波段组合,分别采用基于像素支持向量机分类方法和面向对象影像分类方法对红树林物种进行分类。试验结果表明,基于像素支持向量机分类方法的总体精度为81.03%;利用面向对象影像分类方法的总体精度为85.58%。面向对象影像分类方法能有效去除椒盐噪声,充分利用对象光谱、形状及纹理信息,提供更准确的红树林分布信息。     

New Triangle Vegetation Indices for Estimating Leaf Area Index on Maize

Vegetation index-based methods have been widely used to determine the leaf area index (LAI). Nevertheless, under the high canopy coverage, the estimation ability of current inversion models has been profoundly decreased, due to the “saturation” phenomenon. In this study, the LAI of maize was investigated under various growth conditions. Two new triangular vegetation indices were proposed to improve the inversion ability and estimation accuracy of LAI on maize. The triangle difference vegetation index (TDVI) and triangle ratio vegetation index (TRVI) were constructed, and their accuracies were compared with the present spectral vegetation index models. The result shows that TDVI and TRVI are highly linearly correlated with LAI. The coefficients of determination (R²) and root-mean-square errors are, respectively, 0.92 and 0.94, and 1.42 and 0.92 using the simulated data, while they are, respectively, 0.83 and 0.77, and 0.98 and 1.05 using the measured data. In comparison with other vegetation indices (e.g. MSR, MTVI2, RTVI), TDVI is better able to estimate the LAI of maize. Conversely, TRVI has better inversion ability when the LAI is more than 3. Overall, TDVI is an accurate and robust approach for estimating the LAI of maize. The proposed TDVI and TRVI can be jointly used to retrieve LAI at various canopy coverages.     

Seasonal trend analysis (STA) of MODIS vegetation index time series for the mangrove canopy of the Teacapan-Agua Brava lagoon system,Mexico

Monthly time series, from 2001 to 2016, of the Normalized Difference Vegetation Index (NDVI) and the Enhanced Vegetation Index (EVI) from MOD13Q1 products were analyzed with Seasonal Trend Analysis (STA), assessing seasonal and long-term changes in the mangrove canopy of the Teacapan-Agua Brava lagoon system, the largest mangrove ecosystem in the Mexican Pacific coast. Profiles from both vegetation indices described similar phenological trends, but the EVI was more sensitive in detecting intra-annual changes. We identified a seasonal cycle dominated by Laguncularia racemosa and Rhizophora mangle mixed patches, with the more closed canopy occurring in the early autumn, and the maximum opening in the dry season. Mangrove patches dominated by Avicennia germinans displayed seasonal peaks in the winter. Curves fitted for the seasonal vegetation indices were better correlated with accumulated precipitation and solar radiation among the assessed climate variables (Pearson’s correlation coefficients, estimated for most of the variables, were r ≥ 0.58 p < 0.0001), driving seasonality for tidal basins with mangroves dominated by L. racemosa and R. mangle. For tidal basins dominated by A. germinans, the maximum and minimum temperatures and monthly precipitation fit better seasonally with the vegetation indices (r ≥ 0.58, p < 0.0001). Significant mangrove canopy reductions were identified in all the analyzed tidal basins (z values for the Mann-Kendall test ≤ ?1.96), but positive change trends were recorded in four of the basins, while most of the mangrove canopy (approximately 87%) displayed only seasonal canopy changes or canopy recovery (z > ?1.96). The most resilient mangrove forests were distributed in tidal basins dominated by L. racemosa and R. mangle (Mann-Kendal Tau t ≥ 0.4, p ≤ 0.03), while basins dominated by A. germinans showed the most evidence of disturbance.     

基于Sentinel-2的潮间红树林提取方法

位于潮间带的红树林可能在高潮时被海水淹没的特点,使得传统的植被提取方法在红树林信息提取方面存在局限性。本文在对比分析了出露的红树林、高潮水位淹没的红树林、海水水体的光谱特征后,提出了一种利用归一化潮间红树林指数（NIMI）提取潮间带红树林的方法。该指数是由植被强吸收的红波段,强反射的两个红边波段和近红外波段组成的归一化表达式。利用该指数对福建省龙海九龙江口湿地的红树林进行了分类提取,提取结果与高分二号影像目视验证和现场调查结果进行了对照。结果显示,该方法提取红树林的用户精度达到93.98%,并显著优于利用归一化水体指数（NDWI）、归一化植被指数（NDVI）及随机森林的结果。     

融合Landsat 8与Sentinel-2数据的红树林物候信息提取与分类

科学准确地监测红树林是保护海陆过渡性生态系统的基础和前提,但红树林分布于潮间带,难以进行大规模人工监测。遥感技术能够对红树林进行长时间、大面积监测,但已有研究尚存不足。一方面,红树林分布于热带、亚热带区域,受到天气条件限制难以获得长时间覆盖的有效光学遥感数据;另一方面,红树林极易与其他陆生植被混淆,仅利用多波段数据的光谱信息难以精确识别。本文以恒河三角洲孙德尔本斯地区为例,基于谷歌地球引擎GEE（Google Earth Engine）获取2016年全年的Landsat 8 OLI和Sentinel-2 MSI数据,利用物候信息进行红树林提取研究。首先,基于最小二乘回归构建两个传感器在相同指数之间的关系,重建时间序列数据,之后根据可分性判据选取增强型植被指数EVI（Enhanced Vegetation Index）和陆地表面水分指数LSWI（Land Surface Water Index）。其次,对两个指数的时间序列数据进行Savitzky-Golay滤波处理,并分别提取生长期始期等13种物候信息。最后,将两个指数的物候信息进行特征级联,采用随机森林RF（Random Forest）方法进行分类,提取研究区红树林范围。实验结果表明：Landsat 8 OLI和Sentinel-2 MSI数据融合可有效提升时间序列质量,与基于单一传感器数据的分类结果相比,总体精度提高1.58%;物候信息可以显著增强红树林与其他植被的可分性,与直接使用时间序列数据的分类结果相比,总体精度提高1.92%;同时考虑EVI和LSWI指数可极大地提升分类效果,与采用单一指数相比,总体精度分别提高14.11%和9.69%。因此,本文通过数据融合、物候信息提取和指数特征级联可以更好地提取红树林,总体精度达到91.02%,Kappa系数为0.892。研究验证了物候信息在红树林遥感监测中的应用潜力,提出的方法对科学准确地监测全球或区域红树林具有一定参考价值。     

冠层特征尺度的定量计算模型与方法

冠层特征尺度是植被定量遥感的基础概念,其物理定义和数学定量表达具有重要的研究意义。首先,基于光学辐射传输角度提出的冠层特征尺度的物理定义,即水平维线性混合条件下的最小分辨率单元,建立了冠层特征尺度的数学计算模型,并引入倒置的地统计学指数模型。然后,提出了基于局部方差分析的冠层特征尺度计算方法。最后,利用森林区域高分辨率图像,对论文提出的冠层特征尺度模型进行了定量验证。结果表明,冠层特征尺度模型计算的冠层特征尺度与树林株行距存在密切联系,线性复相关系数达0.95,证明了本文方法的合理性和可行性。本文提出的冠层特征尺度模型为地表特征尺度定量计算提供了一种新方法。     

红树林湿地植被生物量的雷达遥感估算

根据雷达后向散射系数建立了红树林湿地植被生物量的估算模型,并运用遗传算法确定其中非线性模型的最优参数.对比分析表明,雷达后向散射系数模型比NDVI模型在植被生物量估算中有更高的精度.使用NDVI指数有可能导致某些植被类型的生物量估算出现较大的误差.这是因为一些具有密集冠层的草本植被（例如互花米草等）有比红树林高得多的NDVI值.而雷达遥感所具有的侧视特点及一定的穿透能力能有效地获取植被的垂直信息,大大减低植被生物量估算的误差.     

高光谱反演水稻叶面积指数的主成分分析法

为了通过水稻冠层反射光谱来提取水稻叶面积指数信息，尝试利用辐射传输模型PROSPECT+SAIL来模拟水稻冠层反射光谱， 比较了各植被指数中叶面积指数(LAI)和叶绿素浓度的相关性。在观察光谱曲线后发现，红边位置光谱可以较好地区分LAI和叶绿素 浓度二者引起光谱变化的差异。由此提出对700 nm~750 nm区间内的反射光谱做主成分变换，并利用第2主成分与LAI建立反演模型( 即主成分分析法)，取得了较好效果，表明在植被指数趋近于饱和以至于无法区分二者相关性时，主成分分析法可以作为一种简单 而有效提取水稻叶面积指数信息的补充手段。     

利用Sentinel-2影像超分辨率重建的红树林冠层氮含量反演

氮素是植被整个生命周期的必要元素,红树林冠层氮素含量（CNC）遥感估算对红树林健康监测具有重要意义。以广东湛江高桥红树林保护区为研究区,本文旨在基于Sentinel-2影像超分辨率重建技术进行红树林CNC估算和空间制图。研究首先基于三次卷积重采样、Sen2Res和SupReMe算法实现Sentinel-2影像从20 m分辨率到10 m的重建;然后以重建后的影像和原始20 m影像为数据源构建40个相关植被指数,采用递归特征消除法（SVM-RFE）确定CNC估算的最优变量组合,进而构建CNC反演的核岭回归（KRR）模型;最后选取最优模型实现CNC制图。研究结果表明：基于Sen2Res和SupReMe超分辨率算法的重建影像不仅与原始影像具有很高的光谱一致性,且明显提高了影像的清晰度和空间细节。红树林CNC反演波段主要集中在红（B4）、红边（B5）、近红外波段（B8a）以及短波红外波段（B11和B12）,与“红边波段”相关的植被指数（RSSI和TCARIre1/OSAVI）也是红树林CNC反演的有效变量。基于3种方法重建后10 m的影像构建的模型反演精度（R²val>0.579）均优于原始20 m的影像（R²val=0.504）;基于Sen2Res算法重建影像构建的反演模型拟合精度（R²val=0.630,RMSE_val=5.133,RE_val=0.179）与基于三次卷积重采样重建影像的模型拟合精度（R²val=0.640,RMSE_val=5.064,RE_val=0.179）基本相当,前者模型验证精度（R²cv=0.497,RMSE_cv=5.985,RE_cv=0.214）较高且模型变量选择数量最为合理。综合重建影像光谱细节及模型精度,基于Sen2Res算法重建的Sentinel-2影像在红树林CNC估算中具有良好的应用潜力,能为区域尺度红树林冠层健康状况的精细监测提供有效的方法借鉴和数据支撑。     

Roles of horizontal and vertical tree canopy structure in mitigating daytime and nighttime urban heat island effects

The urban heat island (UHI) is increasingly recognized as a serious, worldwide problem because of urbanization and climate change. Urban vegetation is capable of alleviating UHI and improving urban environment by shading together with evapotranspiration. While the impacts of abundance and spatial configuration of vegetation on land surface temperature (LST) have been widely examined, very little attention has been paid to the role of vertical structure of vegetation in regulating LST. In this study, we investigated the relationships between horizontal/vertical structure characteristics of urban tree canopy and LST as well as diurnal divergence in Nanjing City, China, with the help of high resolution vegetation map, Light Detection and Ranging (LiDAR) data and various statistical analysis methods. The results indicated that composition, configuration and vertical structure of tree canopy were all significantly related to both daytime LST and nighttime LST. Tree canopy showed stronger influence on LST during the day than at night. Note that the contribution of composition of tree canopy to explaining spatial heterogeneity of LST, regardless of day and night, was the highest, followed by vertical structure and configuration. Combining composition, configuration and vertical structure of tree canopy can take advantage of their respective advantages, and best explain variation in both daytime LST and nighttime LST. As for the independent importance of factors affecting spatial variation of LST, percent cover of tree canopy (PLAND), mean tree canopy height (TH_Mean), amplitude of tree canopy height (TA) and patch cohesion index (COHESION) were the most influential during the day, while the most important variables were PLAND, maximum height of tree canopy (TH_Max), variance of tree canopy height (TH_SD) and COHESION at night. This research extends our understanding of the impacts of urban trees on the UHI effect from the horizontal to three-dimensional space. In addition, it may offer sustainable and effective strategies for urban designers and planners to cope with increasing temperature.     

引黄灌区水稻叶面积指数的高光谱估测模型

水稻叶面积指数（leaf area index,LAI）是评价其长势的重要农学参数,高光谱遥感能够实现叶面积指数的快速无损监测。为了寻找反演水稻LAI的最优植被指数,扩展水稻LAI高光谱估测模型的普适性,选取宁夏引黄灌区水稻为研究对象,通过设置不同氮素处理,借助相关分析、回归分析等方法研究高光谱植被指数与水稻LAI之间的定量关系,并通过确立的最优波段组合,构建4种植被指数与水稻LAI的高光谱反演模型。结果表明,水稻LAI在抽穗末期达到最大值,并随氮素水平的增加而增加;水稻冠层原始光谱反射率在400~722 nm和1 990~2 090 nm波段与LAI达到极显著负相关水平,在近红外区域760~1 315 nm与LAI呈极显著正相关。模型检验结果表明,以比值植被指数RVI（_850,750）为变量建立的水稻LAI估测模型最佳,研究结果可为水稻LAI的高光谱估测提供地域参考。     

简缩极化特征值分析的溢油检测

为有效利用简缩极化SAR进行海洋溢油检测,本文基于简缩极化特征值分析,提出了3个用于简缩极化溢油检测的参数,引入了基于简缩极化特征值分解的简缩极化熵Hc(Compact Polarization Entropy)、简缩极化比参数PFc(Compact Polarization Fraction)、简缩极化基准高度PHc(Compact Polarization Pedestal Height)特征进行海洋溢油检测。海表的散射类型主要为低熵散射(小粗糙面发生的Bragg散射),为弱去极化、弱散射过程随机性状态,由于溢油会阻尼海水的Bragg散射,使其熵值变高、呈去极化、强散射过程随机性状态,故简缩极化熵、简缩极化比参数和简缩极化基准高度可以用来检测海洋溢油。本文采用C波段的Radarsat-2、SIR-C/X-SAR数据进行了实验,结果表明:简缩极化熵、简缩极化比参数和简缩极化基准高度能够有效抑制疑似溢油,使海水与疑似溢油差异变小;突出溢油区域,使海水与溢油的可区分性变大。     

SAR图像不同油膜特征的最佳分割方法选择

SAR图像溢油分割是SAR溢油监测中一个重要环节。文中选取4种不同形状、尺寸和对比度的SAR油膜数据,分别采用双峰阈值分割法、最大熵分割法、区域生长法、分水岭算法、图割法、水平集方法等6种方法进行溢油信息提取,探讨适合于不同油膜特征的最佳提取方法。结合现有尺度分割标准,提出一种SAR图像溢油信息评价指数——有效分割指数(Effective Segmentation Index,ESI),对不同分割方法得到的溢油提取结果进行定量评价,得出了不同特征油膜所适合的最佳分割方法。     

基于地面试验的植被覆盖率估算模型及其影响因素研究

以植被覆盖率的遥感反演为研究主线,以玉米作物为例,在基于地面试验获得作物光谱、叶面积指数和多角度覆盖率的基础上,对目前普遍采用的两种基于植被指数的植被覆盖率估算模型进行了精度比较,同时对植被覆盖率反演的影响因子(叶面积指数、植被空间分布和观测角度)进行了分析.由此得到:估算植被覆盖率的最优植被指数为归一化植被指数;叶面积指数对植被指数与植被覆盖率间关系的影响随植被的生长不断增大;植被空间分布对垂直覆盖率的估算影响很小.对于多角度覆盖率有这样的规律,即在4种空间分布下,以0°观测天顶角(VZA)为中心,在相反方位角上随VZA的增加,覆盖率值基本呈对称分布;在玉米刚出苗时,覆盖率随VZA的增加而增加,当VZA=0°时达到最小值,而随着玉米的进一步生长,4种分布条件下覆盖率随VZA的增加反而降低,在VZA=0°时达到最大值.     

植被冠层立体结构与叶片倾角的偏振光效应

植被偏振特性研究对于植被监测与组分定量反演具有极其重要的作用。植被冠层的反射辐射具有偏振特性,这种特性与入射辐射和植被冠层结构相关。本文分析了偏振对光子—叶片—冠层之间细微相互作用及其变化的有效探测能力,并利用研究型扫描式偏振辐射仪RSP(Research Scanning Polarimeter)数据系统对比分析了偏振对不同叶倾角分布的估测。通过上述研究得出以下结论:(1)偏振观测能够对光线在冠层立体结构中的透射反射再出射过程给出精细刻画,若不用偏振手段对这一过程进行甄别并去除,则直接测算的植被散射系数会产生高达140%的误差;(2)利用偏振手段可以为高精度大倾角、多时相遥感观测提供可能,以此可改变目前光学遥感小角度、垂直观测的较严格约束;(3)偏振辐射呈现出随波长的稳定特性(相关系数0.96),使得利用偏振手段可以更好地研究冠层结构;(4)不同叶倾角分布对入射辐射存在不同的偏振反射,为利用多角度偏振信息进行遥感植被精细分类提供了新的途径。本文详细描述冠层结构和植被偏振特性的相互作用,通过对冠层立体结构与叶倾角的研究,刻画了植被定量遥感的方向性信息与高精度实现,为高分辨率遥感定量化的有效信息挖掘提供了新手段。     

Retrieval of remotely sensed LAI using Landsat ETM+ data and ground measurements of solar radiation and vegetation structure: Implication of leaf inclination angle

A time series of leaf area index (LAI) of a managed birch forest in Germany (near Dresden) has been developed based on 16-day normalized difference vegetation index (NDVI) data from the Landsat ETM+ sensor at 30 m resolution. The Landsat ETM+ LAI was retrieved using a modified physical radiative transfer (RTM) model which establishes a relationship between LAI, fractional vegetation cover (fC), and given patterns of surface reflectance, view-illumination conditions and optical properties of vegetation. In situ measurements of photosynthetically active radiation (PAR) and vegetation structure parameters using hemispherical photography (HSP) served for calibration of model parameters, while data from litter collection at the study site provided the ground-based estimates of LAI for validation of modelling results. Influence of view-illumination conditions on optical properties of canopy was simulated by a view angle geometry model incorporating the solar zenith angle and the sensor viewing angle. Effects of intra-annual and inter-annual variability of structural properties of the canopy on the light extinction coefficient were simulated by implementing variability of the leaf inclination angle (LIA), which was confirmed in the study site. The results revealed good compatibility of the produced Landsat ETM+ LAI data set with the litter-estimated LAI. The results also showed high sensitivity of the LAI retrieval algorithm to variability of structural properties of the canopy: the implementation of LIA dynamics into the LAI retrieval algorithm significantly improved the model accuracy.     

Angle effects of vegetation indices and the influence on prediction of SPAD values in soybean and maize

To study the anisotropy of vegetation indices (VIs) and explore its influence on the retrieval accuracy of canopy soil-plant analyzer development (SPAD) value, the bidirectional reflectance distribution function (BRDF) models of soybean and maize are calculated from the multi-angle hyperspectral images acquired by UAV, respectively. According to the reflectance extracted from the BRDF model, the dependences of 16 commonly-used VIs on observation angles are analyzed, and the SPAD values of maize and soybean canopy are predicted by using the 16 VI values at different observation angles and their combinations as input parameters. The results show that the 16 VIs have different sensitivity to angle in the principal plane: green ratio vegetation index (GRVI), ratio vegetation index (RVI), red edge chlorophyll index (CIRE), and modified chlorophyll absorption in reflectance index/optimized soil-adjusted vegetation index (MCARI/OSAVI) are very sensitive to angles, among which MCARI/OSAVI of maize fluctuated the most (138.83 %); in contrast, the green optimal soil adjusted vegetation index (GOSAVI), normalized difference vegetation index (NDVI), and green normalized difference vegetation index (GNDVI) hardly change with the observation angles. In terms of SPAD prediction, the accuracy of different VI is different, the mean absolute error (MAE) showed that MCARI1 provided the highest accuracy of retrieval for soybean (MAE=1.617), while for maize it was MCARI/OSAVI (MAE=2.422). However, when using the same VI, there was no significant difference in the accuracy of the predicted results, whether the VI from different angles was used or the combination of multi-angles was used. The present results provide guiding significance and practical value for the retrieval of SPAD value in vegetation canopies and in-depth applications of multi-angular remote sensing.     

Vegetation图像植被指数与实测水稻叶面积指数的关系

水稻的叶面积指数 (LAI)是水稻生长的一项重要参数 ,与水稻的生物量与产量直接相关。利用 1999年在江苏省江宁县实测的水稻叶面积指数与同期Vegetation/SPOT的植被指数作了对比分析 ,结果发现同期的LAI与植被指数表现相近的变化特征 ,两者具有良好的相关关系。     

基于TM的辐射传输模型反演叶面积指数可行性研究

基于PROSAIL辐射传输模型，引入土壤反射指数SRI来简化模型，提出直接从反射率计算SRI的方法；  同时，针对不同的植被状况，采取不同波段组合对模型的参数进行敏感性分析，确定自由参数与反演波段组合，提出一种基于不同植被状况的叶面积指数反演策略； 最后，应用遗传算法对模拟的TM光谱反射数据进行实验。结果表明，对于LAI＜3的植被，反演精度较高； 但是对于LAI＞3的植被，反演精度较低，其原因主要是冠层反射对LAI不再敏感。因此，辐射传输模型反演LAI有一定适用范围，只有在此范围内LAI的反演精度才可靠。