东中国海表层悬浮体浓度卫星遥感反演研究进展

用卫星遥感手段反演海洋表层悬浮体浓度(Suspended Sediment Concentration,SSC)来研究其分布和输运的方法已经被广泛使用。东中国海属于水文和光学性质较为复杂的二类水体,表层悬浮体浓度的分布规律和水体的固有光学特性时空变化大,增加了遥感研究的难度。在对前人的研究进行比较和总结后发现,根据实测SSC数据对不同区域、不同时间段(季节、潮汐周期)建立分段模型可以提高整体反演精度。在选择参与反演的波段时,河口和近岸等高SSC海域以及远岸低SSC海域有各自不同的最优波段组合。高SSC海域常使用水体反射率第二反射峰、第一反射峰前波段作正比波段组合参与反演,低SSC海域常使用水体反射率第一峰波段作正比、峰前波段作反比参与反演。同时,在反演模型中考虑泥沙粒径的影响可以显著提升反演精度,并且也有可能在浅海区突破现有遥感研究手段的水深限制。目前模型精度评价标准使用较为混乱,平均相对误差、平均绝对误差和均方根误差等可以作为综合精度评价指标,模型的稳定性则可以用误差敏感性分析方法验证。高时空分辨率的海色卫星传感器的出现使得海洋短时间尺度事件的研究成为海色遥感研究的趋势之一。     

基于水体光学分类的鄱阳湖悬沙浓度反演方法研究

鄱阳湖是我国重要的湿地生态系统,对调节流域的水沙变化有着重要作用。由于鄱阳湖湖区面积广,内部差异大,单一的悬浮泥沙反演模型不足以准确反演出湖区的悬浮泥沙浓度。以实测的反射光谱数据、泥沙浓度数据为基础,提出一种基于分类后的反演模型,即根据实测数据的光谱形态特征分类出5种典型的水体类型。在此基础上,将分类后的各类水体分别建立各自合适的反演模型进行反演。结果表明基于水体分类的经验模型反演达到满意效果,平均绝对误差为0.00217g/L,平均相对误差为3.022%。基于分类后的经验反演模型适用于鄱阳湖悬沙浓度分布的监测研究,有助于更加宏观、准确的掌握鄱阳湖泥沙浓度的空间分布和变化,为保持鄱阳湖资源的可持续开发与利用提供决策依据。     

利用MODIS反演长江中游悬浮泥沙含量的初步研究

在众多卫星传感器中,中等分辨率成像光谱仪(MODIS)数据因其高的时间分辨率和中等的空间分辨率,对于水质研究具有自身的潜力和优越性.选取长江中游主河道武汉至宜昌段为例,利用MODIS 250 m波段数据定量反演了内陆河流悬浮泥沙的质量浓度.研究结果表明,地面实测的悬浮泥沙质量浓度与MODIS1,2波段的反射率组合(R1-R2)/(R1 R2)有很好的相关关系(相关系数 R2=0.72,样品数n=41),基于这种相关性建立了长江中游主河道武汉至宜昌段表层悬浮泥沙的遥感定量反演经验模型.     

基于遥感方法反演悬浮泥沙分布

基于遥感理论,采用地物光谱仪与光电测沙仪对同一水样同步测量的方法,确保建立相关关系数据的同步性;由地物光谱仪测得的光谱值推算出遥感反射率,并等效计算ETM₊图象相应波段的遥感反射率;由光电测沙仪测出悬浮泥沙浓度;通过对等效算出的ETM₊各个波段遥感反射率和悬浮泥沙浓度这两组数据的相关分析,得到4种形式的经验模型.结果表明,TM₄与TM₁波段光谱值的比值与悬浮泥沙浓度之间的相关系数最高,故应用该模式进行研究区域悬沙浓度反演.     

基于高光谱数据提取水体叶绿素a浓度的混合光谱模型

从模型的精度和稳定性方面,与传统的经验模型比较显示,利用原始数据的一阶微分数据进行叶绿素a浓度混合光谱模型反演最佳。最后,以此模型对三种不同年份测量的地面高光谱数据进行了叶绿素a浓度的提取。实验结果表明混合光谱模型可以作为遥感监测水体叶绿素a含量的定量模型。     

基于MODIS数据的云南抚仙湖星云湖水质污染遥感调查方法研究

水质遥感监测方法可以反映水质在空间和时间上的分布情况和变化,满足目前对水质调查实时、大尺度的评价要求。文章以云南抚仙湖星云湖为研究区,将实测的叶绿素a浓度、悬浮物含量和MODIS遥感数据各波段比值进行相关分析,结果表明:MODIS数据250m分辨率的r2/r1的波段组合能够较好地反演叶绿素a浓度,而500m分辨率的(r4-r3)/(r4+r3)的波段组合对于悬浮物浓度的反演能力最好。通过对它们进行回归关系分析,得到适合抚仙湖星云湖的遥感定量模型。     

基于遗传神经网络的克钦湖叶绿素反演研究

叶绿素a浓度能够在一定程度上反映内陆湖泊水质情况。为实现对克钦湖水体叶绿素a浓度的监测,于2010年8月15日对克钦湖进行了现场光谱测量和同步采样。通过分析叶绿素a浓度和光谱数据之间的关系,建立基于反射比、人工神经网络和遗传神经网络的叶绿素a浓度估测模型。结果表明:利用R700nm/R670nm反射比建立的模型估测精度为R2=0.67;人工神经网络模型的估测精度较高,R2=0.882;将遗传算法引入神经网络之后,模型的估测精度进一步提高,R2达到0.956,将模型预测的结果与克里格内插法相结合对研究区的叶绿素a空间分布情况进行定量估测,发现北湖的叶绿素a浓度明显高于南湖,有由北向南逐渐递减的趋势,这为今后利用高光谱数据对克钦湖叶绿素a浓度大面积遥感反演提供了研究基础。     

基于时空融合算法的水体叶绿素a反演研究

为了准确反演水体中叶绿素a浓度,以黄柏河东支流域为例,采用STNLFFM时空融合算法,对2017年GF-4和Sentinel-2反射率数据进行融合,以重构Sentinel-2影像的时间序列数据,并对应用算法前后获取的水质参数-光谱特征响应关系建立多元线性回归模型,比较模型对叶绿素a的预测效果以验证时空融合算法的可行性,利用重构后影像光谱特征与水质参数的响应关系建立人工神经网络模型,反演2017年黄柏河东支流域各水库水体叶绿素a浓度。结果表明：利用时空融合算法生成的影像接近真实影像,提高了多元线性回归模型预测叶绿素a的效果,R2从融合前0.659提高至融合后0.844,且基于时空融合算法获取的水质参数-光谱关系建立的人工神经网络模型模拟精度较好,R2和MRE达到0.925和9.461%,反演的叶绿素a浓度空间差异性明显。证明了时空融合算法在水质参数反演过程中具有较好的应用前景。     

长春市南湖叶绿素a浓度高光谱估测模型研究

利用野外高光谱仪(ASD Field Spec radiometer)在长春南湖进行了反射光谱测量和同步水质采样分析,通过研究水体藻类叶绿素a浓度与其高光谱反射特征之间的相关关系,采用单波段反射率、反射率比值法和一阶微分法分别建立了叶绿素a的高光谱定量模型。结果表明:三者与叶绿素a线性相关程度都比较高,决定系数在0.70左右,而且其显著水平P<0.01,都可以用于叶绿素a的定量遥感,其中单波段模型效果最好。为今后利用高光谱传感器在南湖进行叶绿素a浓度大面积遥感反演提供了研究基础。     

应用MODIS影像估测太湖水体悬浮物浓度

以太湖为研究区域,同步获取悬浮物浓度实测数据、水体反射光谱数据和MODIS卫星影像数据,构建基于中分辨率成像光谱仪(MODIS)的悬浮物遥感估测模型.为了削弱大气效应,对MODIS影像了进行了粗略大气纠正.通过悬浮物特征光谱分析,将MODIS各敏感波段及波段组合与悬浮物浓度实测值进行相关分析,并应用实测光谱数据进行验证.在此基础上,运用回归分析建立半经验反演模型,并对模型进行了评价和应用.研究结果表明,MODIS影像可以很好地对大型内陆湖泊的悬浮物浓度进行遥感估测.250 m波段2 500 m波段4与1 000 m波段14是探测悬浮物的敏感波段.波段组合上,500 m组合因子r4/r3、r4-r3估测悬浮物含量的精度很高,适于构建反演模型;1 000 m波段8、11、131、4的多元组合也是构建模型的较好选择(R2均不低于0.85).     

Spatial and temporal variations of suspended solid concentrations from 2000 to 2013 in Poyang Lake,China

Concentrations of suspended solids in lakes can affect the latter’s primary productivity and reflect changes in sediment deposition. Determining the temporal and spatial distribution of suspended solid concentrations has important significance in lake water environmental management; this is particularly urgent for Poyang Lake, the largest freshwater lake in China. In this study, suspended solid concentration inversion models for Poyang Lake were created using a semi-empirical method with regression analysis between continuously measured suspended solid concentration data and multi-band moderate-resolution imaging spectroradiometer images for spring, summer, autumn, and winter from 2009 to 2012. The coefficient of determination (R²) is from 0.6 to 0.9 and the average relative error for the accuracy verification was between 10 and 30%. The seasonal distributions of suspended solid concentrations in Poyang Lake from 2000 to 2013 were then obtained using optimal reversal models. The results showed that the seasonal variation in suspended solid concentrations had a “W” shape in which high spring and autumn and low summer and winter values. The suspended solid concentrations increased annually from 2000 to 2013 and were mainly distributed in the northern and central portions of the lake, with lower values along the shorelines. Further analysis indicated that the large difference in water level between the wet and dry seasons is an important factor in explaining these seasonal variations. Moreover, the suspended solid concentrations were poorly correlated with water temperature and chlorophyll-a concentration but more highly correlated with the deferred chlorophyll-a concentration.     

A hydrologic regression sediment-yield model for two ungaged watershed outlet stations in Africa

A hydrologic regression sediment-yield model was established to determine the relationship between water discharge and suspended sediment discharge at the Blue Nile and the Atbara River outlet stations during the flood season. The model consisted of two main submodels: (1) a suspended sediment discharge model, which was used to determine suspended sediment discharge for each basin outlet; and (2) a sediment rating model, which related water discharge and suspended sediment discharge for each outlet station. Due to the absence of suspended sediment concentration measurements at or near the outlet stations, a minimum norm solution, which is based on the minimization of the unknowns rather than the residuals, was used to determine the suspended sediment discharges at the stations. In addition, the sediment rating submodel was regressed by using an observation equations procedure. Verification analyses on the model were carried out and the mean percentage errors were found to be +12.59 and –12.39, respectively, for the Blue Nile and Atbara. The hydrologic regression model was found to be most sensitive to the relative weight matrix, moderately sensitive to the mean water discharge ratio, and slightly sensitive to the concentration variation along the River Nile's course.     

基于WorldView-02高分影像的BP和RBF神经网络遥感水深反演

遥感水深反演是水深测量的一种重要技术和手段.以美济礁水深反演为例,选择WorldView-02高分影像为数据源,在辐射定标和大气校正的基础上,构建BP（Back Propagation）和RBF（Radial Basis Function）人工神经网络水深反演模型,以遥感影像8个波段为输入层,通过tansig、logsig、高斯函数和purelin函数变换实现从输入层到隐含层、隐含层到输出层的转换,以便反演水深.最后对反演水深与实测水深采用回归分析,求解决定系数（coefficient of determination,R2）、平均决定误差（Mean Absolute Error,MAE）、均方根误差（Root Mean Square Error,RMSE）等进行比较,评价2种模型的精度.结果表明,RBF神经网络模型结构更简单,对样本要求更低,反演精度达到0.995,更适合遥感水深反演.      

青藏高原Soumi-NPP和MODIS积雪范围产品的对比分析

Soumi-NPP（Soumi Polar-orbiting Partnership）卫星作为接替服役超期的Terra、Aqua卫星,其积雪范围产品在青藏高原的精度尚未被评价。以Soumi-NPP积雪范围产品为研究对象,利用气象台站点数据并结合更高分辨率的Landsat-8 OLI数据,评价该产品的精度,并与MODIS（Moderate Resolution Imaging Spectroradiometer）积雪范围产品进行对比分析。结果表明：使用气象台站进行数据验证时,NPP、MOD与MYD三种积雪范围产品的总精度均较高,但三者积雪漏分误差都较大,其中MYD的漏分误差最大,为64.2%;当雪深小于5 cm时,三种积雪范围产品的积雪分类精度都较低,雪深大于等于5 cm时,NPP积雪范围产品的积雪分类精度最高,为82.3%,MOD与MYD的精度分别为77.1%和69.4%;利用Landsat-8 OLI数据验证时,Soumi-NPP积雪范围产品的Kappa系数最高,其均值为0.707,为高度一致性。而MOD10A1与MYD10A1的Kappa系数较低,分别为0.476与0.557,为中等一致性;Soumi-NPP积雪范围产品的Kappa系数大多在0.6以上,精度比较稳定,而MODIS积雪范围产品的Kappa系数波动较大,精度稳定性较差。Soumi-NPP积雪范围产品相较于MODIS积雪范围产品,其精度有了较大的提升,为准确监测青藏高原积雪范围提供了一个更优的选择。     

Comparison of cokriging and adaptive neuro-fuzzy inference system models for suspended sediment load forecasting

The present research was carried out by using artificial neural network (ANN), adaptive neuro-fuzzy inference system (ANFIS), cokriging (CK) and ordinary kriging (OK) using the rainfall and streamflow data for suspended sediment load forecasting. For this reason, the time series of daily rainfall (mm), streamflow (m³/s), and suspended sediment load (tons/day) data were used from the Kojor forest watershed near the Caspian Sea between 28 October 2007 and 21 September 2010 (776 days). Root mean square error, efficiency coefficient, mean absolute error, and mean relative error statistics are used for evaluating the accuracy of the ANN, ANFIS, CK, and OK models. In the first part of the study, various combinations of current daily rainfall, streamflow and past daily rainfall, streamflow data are used as inputs to the neural network and neuro-fuzzy computing technique so as to estimate current suspended sediment. Also, the accuracy of the ANN and ANFIS models are compared together in suspended sediment load forecasting. Comparison results reveal that the ANFIS model provided better estimation than the ANN model. In the second part of the study, the ANN and ANFIS models are compared with OK and CK. The comparison results reveal that CK was a better estimation than the OK. The ANFIS and ANN models also provided better estimation than the OK and CK models.     

基于Worldview-3与Sentinel-1 SAR数据的草原矿区复垦植被生物量反演方法研究

利用多源遥感数据定量反演矿区复垦植被生物量是高效、动态、大面积监测土地复垦和生态恢复效果的必要手段之一。本文以内蒙古草原露天煤矿为研究区,联合遥感光学与雷达数据各自的优势,探索基于Worldview-3(WV-3)与Sentinel-1 SAR数据的矿区复垦植被生物量反演方法,选择主成分-小波变换(W-PCA)算法对WV-3与Sentinel-1 SAR数据进行融合,揭示波段反射率、植被指数、后向散射系数及纹理特征等参数与生物量之间的相关关系,建立多变量的生物量反演模型,并分析不同生物量模型的空间不确定性。结果表明:(1)通过W-PCA算法得到融合后的图像,信息熵的提高反映了融合图像与光学WV-3图像相比具有更多的细节信息,平均梯度的提高反映了融合图像与Sentinel-1 SAR图像相比具有更高的清晰度和更丰富的纹理信息。融合后的第8波段相关系数最高、光谱扭曲度最低、光谱保真度最高。(2)通过相关性分析,发现增强型植被指数(EVI)、归一化植被指数(NDVI)、VH极化、VH均值纹理以及融合后第8波段与生物量显著正相关。WV-3的NDVI与Sentinel-1的VH_ME建模精度(R²=0.834 0,RMSE=16.464 6 g/m²,Ac=81.52%)最高,融合后的第8波段验证精度(R²=0.798 3,RMSE=22.828 3 g/m²,Ac=74.64%)最高。(3)基于不同模型的残差不确定性分析,Sentinel-1 SAR数据变量建立的模型估测结果更容易出现高估及饱和现象,两者联合变量建立的模型可实现优势互补,利用融合数据建立的模型明显改善生物量小于40 g/m²时的高估计现象以及生物量大于100 g/m²时的两者饱和现象,其不确定性降低2.42~9.68 g/m²。因此,利用光学和雷达遥感融合能够有效提高复垦植被生物量的估算精度,为草原矿区复垦植被精细监测提供有效的数据支持。     

Comparison of various pan-sharpening methods using Quickbird-2 and Landsat-8 imagery

Pan-sharpening is the process of transferring the spatial resolution of panchromatic (PAN) image to a multispectral (MS) image for producing a single image with high spatial detail and rich spectral information. In this study, PAN and MS imagery of Quickbird-2 and Landsat-8 are fused separately, using ten different pan-sharpening methods such as principal component analysis (PCA), modified-intensity hue saturation (M-IHS), multiplicative, brovey transform (BT), wavelet-principal component analysis (W-PCA), hyperspectral color space (HCS), high-pass filter (HPF), Gram-Schmidt (GS), Fuze Go, and non-subsampled contourlet transform (NSCT). The effectiveness of these techniques is assessed and compared by qualitative analysis and 14 quantitative analysis methods including bias, correlation coefficient (CC), difference in variance (DIV), relative dimensionless global error in synthesis (ERGAS), universal image quality index (Q), relative average spectral error (RASE), root mean square error (RMSE), structural similarity index method (SSIM), signal-to-noise ratio (SNR), peak SNR (PSNR), spatial correlation coefficient (SCC), image entropy (E), and gradient and quality with no reference image (QNR). The results of both analysis types show that the Fuze Go and NSCT produced the best fused image with high spatial detail and rich spectral information followed by the HPF and GS.     

一种改进的近红外波段悬浮物生物光学反演模型

在分析固有光学参数和表观光学参数之间关系的基础上,提出一种提高悬浮物浓度的反演精度的算法。本文算法细化单位散射系数曲线的斜率s’值,并将bbp（λ）×bρ（λ）视为一个一定数值范围内的变量,其中b’bp（λ）为波长A处的悬浮物后向散射概率,bρ（λ）为悬浮物的单位散射系数。以太湖为实验区,进行了水体野外实测,获得了遥感反射率、吸收散射系数、后向散射系数、悬浮物浓度等必要参数。利用本文改进算法反演悬浮物浓度的平均相对误差绝对值为22．58％,均方根误差为13．12mg／L,均小于已有研究中将涉及的相关参数看做单一定值时,取得的平均绝对误差（44．78％）和均方根误差（31．57mg／L）。