不同森林植被的高光谱特征分析

“生态水(层)”富水特征特殊,各信息指标参数难以用常规方法进行量化和反演,高光谱遥感由于其波段多、光谱信息丰富的优点为生态水(层)各信息指标参数的量化反演提供有效的数据源及方法。利用高光谱遥感技术进行植被分析时,其光谱特征的分析和敏感波段提取非常重要。针对“生态水”信息指标植被参数有关量化反演需要,对研究区部分典型植被叶片进行了光谱采集,利用微分方法对光谱数据进行处理,分析了不同植被叶片光谱的原始、一阶微分和二阶微分光谱曲线,从中提取差异大的波段区分不同植被。同时,采用距离统计分析方法对所选择的不同波段进行有效性验证。研究结果表明:虽然3种方法提取的波段有差异,但存在共同点;选择的光谱特征波段可有效地区分不同植被,在近红外波段尤为明显,分别是1 814~1 823 nm,1 874~1 883 nm和1 890~1 899 nm附近。     

高光谱植被遥感数据光谱特征分析

利用植被的光谱数据,探讨了植被冠层的光谱反射特征和诊断性光谱吸收特征。根据植被光谱特征和连续统去除法(CR),介绍了识别植被种类和预测植被冠层营养元素等生化组分含量的可能性。运用一阶微分反射比(FDR)和从连续统去除的光谱吸收特征中获得的波段深度(BD)、连续统去除后微分反射比(CRDR)、波段深度比(BDR)和归一化波段深度指数(NBDI)等变量,利用逐步线性回归模型并基于光谱吸收特征的变量来选择波长,并通过相关分析来预测植被冠层生化组分。     

基于Hyperion影像的高光谱水汽含量反演的算法比较研究

基于高光谱遥感图像数据的大气参数反演和一体化辐射校正具有重要研究意义和应用价值。首先,通过6S模型辐射传输计算分析了EO-1/Hyperion遥感影像在940和1 130nm附近水汽吸收区域的光谱吸收特点。其次,采用两通道比值法和三通道比值法,比较了不同波段组合的大气含水量高光谱遥感反演精度并进行了敏感性分析,模拟实验结果表明采用三波段比值算法的相关系数和均方根误差均优于对应的两波段算法。最后,利用张掖地区2008年3景EO-1Hyperion高光谱遥感影像,反演了大气含水量,并与地基CE-318太阳分光光度计测量数据进行对比验证,结果表明:1 124nm水汽吸收通道反演精度优于940nm,两通道和三通道比值法的均方根误差分别为0.369和0.128g/cm2,三通道比值方法优于两通道比值方法,与地面观测结果一致。     

基于贝叶斯树算法的高光谱遥感数据分类

文章针对高光谱波段数众多、信息冗余量大的特点,首先对高光谱曲线进行光谱特征参数提取,然后再选择合适的吸收峰波段作为输入向量,在VS2008平台上实现了采用贝叶斯树（NBTree）算法对铀矿床高光谱数据进行分类。     

东北黑土区土壤铬含量高光谱反演研究

以东北黑土区某农场为研究区,在光谱积分定义的基础上,提出一种新的光谱特征参数——反向光谱吸收积分,建立偏最小二乘回归模型对土壤中铬元素含量进行反演研究。与传统的光谱特征参量,包括微分变换、倒数变换、对数变换等11种光谱变换以及吸收面积建立的土壤铬含量高光谱反演模型进行对比分析,结果表明:在光谱变换特征中,平方根一阶微分模型能够较好地定量预测铬含量;吸收面积模型稳定性略差,只能对样本进行初略估计;针对反向光谱吸收积分模型,其建模样本的调整决定系数为0.73,均方根误差为2.63 mg/kg;验证样本的调整决定系数为0.77,均方根误差为2.36 mg/kg,相对偏差为3.21,表明此模型具有极好的预测能力。因此,反向光谱吸收积分能够明显改善铬含量反演模型的精度和稳定性,为土壤铬污染监测提供了新的思路。     

基于KPCA光谱特征约束的水边线提取算法

为充分利用遥感影像的多波段光谱特征,提高水边线的提取精度,提出了基于核主元分析（KPCA）光谱特征约束的水边线提取模型。利用KPCA变换提取水体样本的光谱特征,采用最大似然法估计特征空间中水体光谱特征概率密度函数的特征参数,进而构建水体的光谱特征项。以测地线活动轮廓（GAC）模型为基础,建立图像数据项。结合光谱特征项和图像数据项建立水边线提取模型。在Landsat TM数据集上进行的水边线提取实验验证了算法的有效性,与GAC模型和基于距离正则化的水平集方法（DRLSE）相比较,该算法提取的水边线,在保证一定运行速度的情况下,更准确。     

基于光谱指数的植被含水率遥感反演模型研究——以岷江上游毛尔盖地区为例

利用研究区植被样本实测含水率和实测光谱数据,基于植被光谱指数法,建立植被含水率与植被光谱指数之间的数学模型,同时利用该模型对研究区的遥感数据进行分析,反演植被含水率。结果证明:简单比值光谱指数与植被含水率有较好的相关性,线性模型更适合该研究区的植被含水率反演。1999年和2007年两年的植被含水率反演结果显示:9年间植被含水率提高,含水率高的面积增大。     

基于Sentinel-1与FY-3C数据反演植被覆盖地表土壤水分

基于新一代的Sentinel-1SAR数据与FY-3C的MWRI数据,研究植被覆盖地表土壤湿度反演方法。为消除植被对土壤湿度反演影响,首先利用FY-3C/MWRI的微波极化差异指数MPDI,建立植被含水量反演模型;然后,结合植被含水量反演模型和水—云模型,发展一种主被动微波联合反演植被覆盖地表土壤含水量模型;最后,在江淮地区开展反演试验,利用观测的土壤湿度数据进行反演结果的精度验证。结果表明:(1)对于植被覆盖地表土壤湿度反演,由FY3C/MWRI提取的MPDI对于去除植被影响效果较好;(2)相比于VH极化哨兵1号卫星数据,VV极化数据更适用于土壤含水量的反演,能够得到更高的土壤湿度反演精度;(3)哨兵1号卫星数据能够获得较高精度的土壤含水量反演结果,试验反演的土壤湿度值与实测值相关系数为0.561 2,均方根误差为0.044cm~3/cm~3。     

高光谱数据的光谱信息特点及面向对象的特征参数选择——以北京顺义区为例

波段宽度为纳米级的高光谱数据,具有几十乃至几百个光谱通道,它们各有不同的特点。如何根据具体的应用目的,在这众多的波段中选择出最佳波段和特征参数,对于有效地进行高光谱数据的处理、分析及信息提取至关重要。以北京顺义区高光谱数据为例,首先分析了通道间的相关性,根据通道的相关性大小和分组块状结构特点,将其分为若干组;然后全面分析了高光谱数据的光谱信息特征,在综合考虑各波段的信息含量、波段间的相关性以及地物光谱的吸收特性和可分性等因素
的基础上,提出了面向对象的分层多次选择高光谱数据最佳波段和提取特征参数的基本思路和方法;最后用其它地区的成像光谱数据对此方法进行了验证。     

BP-ANN在荒漠草地高光谱分类研究中的应用

利用高分辨率光谱仪在实地测得的光谱数据来识别新疆阜康地区的7种典型荒漠草种,对原始高光谱数据作预处理（微分和平滑）,选取典型荒漠植被的光谱特征（红边、绿峰、红谷、RVI等）作为输入数据,植被类型作为输出数据,构建基于BP神经网络模型的典型荒漠草地分类器,进行了三组基于高光谱特征的草地类型分类实验,结果表明：（1）红边特征较其余吸收特征更能获得精确的分类结果;（2）波段550~790 nm间的窄波段光谱分类间隔中,20 nm优于10 nm的间隔;（3）草地分类器中BP网络模型的输入层、隐藏层神经元个数与BP网络训练时间、精度具有复杂的耦合关系,不可一概而论。     

Predicting water content using Gaussian model on soil spectra

This paper presents an approach to estimating soil moisture content through fitting an inverted Gaussian function to the continuum in soil spectra. The soil moisture Gaussian model (SMGM) estimates the water content by the declining reflectance in the near infrared (NIR) and shortwave infrared (SWIR) regions, 1.2-2.5 μm, due to the spreading of the fundamental water absorption at 2.8 μm. Convex hull boundary points were used to isolate the spectral continuum and to fit the inverted Gaussian function. The function extrapolates the continuum to the fundamental water absorption beyond the wavelength limits of common laboratory, field, and airborne instruments. Of the derived functional parameters, both amplitude and area on the shortwave side of the inverted Gaussian curve were highly correlated with soil water content.In this study, laboratory spectra, from 0.4 to 2.5 μm, were measured at sequential moisture levels in soil samples collected in Castilla-La Mancha, Spain and in California, USA. The Gaussian area was determined to be the best indicator of gravimetric water content with the initial modeling of 2592 spectra. The SMGM was validated with a separate set of 849 spectra. The model performance significantly improved for water contents below a critical level of 0.32 g water/g soil. Within this restricted range, the SMGM predicted water contents for all soils with a maximum of 0.027 RMSE for 1901 modeled spectra and 0.026 for 602 validation spectra. The water content estimates were improved slightly by stratifying the model and validation sets by the two locations, reducing the RMSE to 0.023 in Spain and 0.025 in California. Further stratifying the model spectra by landform and soil sodicity improved some predictions substantially, but less consistently. Stratifying the samples locally demonstrated that a priori knowledge of soil surfaces by landforms should be part of an image calibration strategy. The SMGM provides practical water content estimates and has a potential use in correcting the effects of soil moisture in hyperspectral images.     

基于偏最小二乘的土壤重金属铜含量高光谱估算

为探究高光谱数据估算土壤重金属铜含量的可行性,以石家庄市水源保护区褐土为研究对象,对不同光谱变换数据与重金属铜含量做了相关分析,建立了土壤重金属铜的单光谱变换指标偏最小二乘模型和多光谱变换指标偏最小二乘模型。结果表明：光谱反射率（R）经倒数一阶微分（RTFD）变换后与铜含量的相关性有所提高;光谱敏感波段为418、427、435、446、490、673、1 909、1 920和2 221 nm,基本位于土壤氧化铁、粘土矿物的特征吸收区域;对土壤重金属铜含量估算效果最好的单光谱变换指标偏最小二乘模型为RTFD模型,其模型决定系数（R²）为0.649,均方根误差（RMSE）为1.477;多光谱变换指标偏最小二乘模型R²和RMSE分别为0.751和1.162,建模效果优于单光谱变换指标模型。研究结果可为北方地区褐土类型土壤重金属铜的高光谱估算提供借鉴。     

基于偏最小二乘的土壤重金属铜含量高光谱估算

为探究高光谱数据估算土壤重金属铜含量的可行性，以石家庄市水源保护区褐土为研究对象，对不同光谱变换数据与重金属铜含量做了相关分析，建立了土壤重金属铜的单光谱变换指标偏最小二乘模型和多光谱变换指标偏最小二乘模型。结果表明：光谱反射率（R）经倒数一阶微分（RTFD）变换后与铜含量的相关性有所提高；光谱敏感波段为418、427、435、446、490、673、1 909、1 920和2 221 nm，基本位于土壤氧化铁、粘土矿物的特征吸收区域；对土壤重金属铜含量估算效果最好的单光谱变换指标偏最小二乘模型为RTFD模型，其模型决定系数（R²）为0.649，均方根误差（RMSE）为1.477；多光谱变换指标偏最小二乘模型R²和RMSE分别为0.751和1.162，建模效果优于单光谱变换指标模型。研究结果可为北方地区褐土类型土壤重金属铜的高光谱估算提供借鉴。     

Design of a feature-tuned ANN model based on bulk rock-derived mineral spectra for endmember classification of a hyperspectral image from an iron ore deposit

In this article, a new feature-tuned artificial neural network (ANN) model has been developed for endmember classification of a hyperspectral image. This model is developed on the basis of using only the essential absorption bands of mineral spectra as opposed to using all the spectral bands of the hyperspectral image. This approach has the added advantages of reducing the dimensionality of input features to the ANN as well as inhibiting the influences of noisy bands for classification of endmembers. The proposed ANN model is trained using input features extracted from laboratory spectra of in situ bulk ore materials collected from an existing iron ore deposit. The input features are basically the constituent absorption bands of mineral spectra where each absorption band is mathematically characterized by the centre, width, and strength parameters of a Gaussian curve. For extracting absorption bands from a mineral spectrum, a modified Gaussian model has been used. The application of this model also necessitates the design of a special template for the input layer ANN model. After training the model, its generalization property is assessed through a testing data set. The model has achieved nearly 97% of classification accuracy in a training set, and 71% of accuracy in a testing set. The trained model is then applied on Hyperion imagery collected over an iron ore deposit. All the endmember spectra of this deposit are classified into either vegetation or any of the ores or rock present in the deposit. None of the endmembers is classified into non-iron ore minerals.     

A novel two-step method for winter wheat-leaf chlorophyll content estimation using a hyperspectral vegetation index

Remote sensing estimation of leaf chlorophyll content is of importance to crop nutrition diagnosis and yield assessment, yet the feasibility and stability of such estimation has not been assessed thoroughly for mixed pixels. This study analyses the influence of spectral mixing on leaf chlorophyll content estimation using canopy spectra simulated by the PROSAIL model and the spectral linear mixture concept. It is observed that the accuracy of leaf chlorophyll content estimation would be degraded for mixed pixels using the well-accepted approach of the combination of transformed chlorophyll absorption index (TCARI) and optimized soil-adjusted vegetation index (OSAVI). A two-step method was thus developed for winter wheat chlorophyll content estimation by taking into consideration the fractional vegetation cover using a look-up-table approach. The two methods were validated using ground spectra, airborne hyperspectral data and leaf chlorophyll content measured the same time over experimental winter wheat fields. Using the two-step method, the leaf chlorophyll content of the open canopy was estimated from the airborne hyperspectral imagery with a root mean square error of 5.18 μg cm^?2, which is an improvement of about 8.9% relative to the accuracy obtained using the TCARI/OSAVI ratio directly. This implies that the method proposed in this study has great potential for hyperspectral applications in agricultural management, particularly for applications before crop canopy closure.     

基于高光谱数据和MODIS影像的鄱阳湖悬浮泥沙浓度估算

本文旨在寻找悬浮泥沙浓度的MODIS遥感影像估算模型,并利用实测的高光谱数据对其敏感波段和反演模型进行测试和验证。以鄱阳湖为研究区域,利用光谱数据进行分析,为利用遥感影像建模提供依据。进一步利用同步进行的鄱阳湖水质采样分析与MODIS影像中等分辨率各个波段反射率及其组合进行相关分析,寻找反演悬浮泥沙浓度的敏感波段。实验表明,MODIS的第一波段反射率对于悬浮泥沙浓度有很好的匹配(R² = 0.91; n = 25),进而建立了鄱阳湖地区的悬浮泥沙浓度遥感定量估算模型。利用估算模型和鄱阳湖地区历史MODIS影像,得到了鄱阳湖悬浮泥沙浓度分布图。基于对汛期鄱阳湖悬浮泥沙浓度的连续监测,可对长江倒灌入鄱阳湖现象的形态进行观测。     

联合无人机激光雷达和高光谱数据反演玉米叶面积指数

叶面积指数（Leaf Area Index, LAI）是作物长势监测及产量估算的重要指标,准确高效的LAI反演对农田经济的宏观管理具有重要作用。研究探索了联合无人机激光雷达（Light Detec-tion and Ranging, LiDAR） 和高光谱数据反演玉米叶面积指数的潜力,并分析了LiDAR数据不同采样尺寸、高度阈值、点密度对LAI反演精度的影响同时确定三者的最优值。该研究分别从重采样的LiDAR数据和高光谱影像中提取了LiDAR变量和植被指数,然后基于偏最小二乘回归（Partial Least Square Regression,PLSR）和随机森林（Random Forest, RF） 回归两种算法分别利用LiDAR变量、植被指数、联合LiDAR变量和植被指数构建预测模型,并确定反演玉米LAI的最优预测模型。结果表明：反演玉米LAI的最优采样尺寸、高度阈值、点密度分别为5.5 m、0.55 m、18 points/m²,研究发现最高的点密度（420 points/m²）并没有产生最优的玉米LAI反演精度,因此单独依靠增加点密度的方法提高LAI的反演精度并不可靠。基于LiDAR变量获得的LAI反演精度（PLSR：R²=0.874,RMSE=0.317;RF：R²=0.942,RMSE=0.222）高于基于植被指数获得的LAI反演精度（PLSR： R²=0.741,RMSE=0.454;RF：R²=0.861,RMSE=0.338）,而使用组合变量构建预测模型的反演精度（PLSR：R²=0.885, RMSE=0.304;RF：R²=0.950,RMSE=0.203）优于使用单一变量建立的LAI预测模型,其中利用联合LiDAR变量和植被指数建立的随机森林回归模型为最优预测模型。因此,将两种数据源融合在提高植被LAI反演精度方面具有一定的潜力。     

Evaluation of hyperspectral data for pasture estimate in the Brazilian Amazon using field and imaging spectrometers

We used two hyperspectral sensors at two different scales to test their potential to estimate biophysical properties of grazed pastures in Rondônia in the Brazilian Amazon. Using a field spectrometer, ten remotely sensed measurements (i.e., two vegetation indices, four fractions of spectral mixture analysis, and four spectral absorption features) were generated for two grass species, Brachiaria brizantha and Brachiaria decumbens. These measures were compared to above ground biomass, live and senesced biomass, and grass canopy water content. The sample size was 69 samples for field grass biophysical data and grass canopy reflectance. Water absorption measures between 1100 and 1250 nm had the highest correlations with above ground biomass, live biomass and canopy water content, while ligno-cellulose absorption measures between 2045 and 2218 nm were the best for estimating senesced biomass. These results suggest possible improvements on estimating grass measures using spectral absorption features derived from hyperspectral sensors. However, relationships were highly influenced by grass species architecture. B. decumbens, a more homogeneous, low growing species, had higher correlations between remotely sensed measures and biomass than B. brizantha, a more heterogeneous, vertically oriented species. The potential of using the Earth Observing-1 Hyperion data for pasture characterization was assessed and validated using field spectrometer and CCD camera data. Hyperion-derived NPV fraction provided better estimates of grass surface fraction compared to fractions generated from convolved ETM+/Landsat 7 data and minimized the problem of spectral ambiguity between NPV and Soil. The results suggest possible improvement of the quality of land-cover maps compared to maps made using multispectral sensors for the Amazon region.     

ICA mixture model algorithm for unsupervised classification of remote sensing imagery

Conventional remote sensing classification algorithms assume that the data in each class can be modelled using a multivariate Gaussian distribution. As this assumption is often not valid in practice, conventional algorithms do not perform well. In this paper, we present an independent component analysis (ICA)‐based approach for unsupervised classification of multi/hyperspectral imagery. ICA used for a mixture model estimates the data density in each class and models class distributions with non‐Gaussian (sub‐ and super‐Gaussian) probability density functions, resulting in the ICA mixture model (ICAMM) algorithm. Independent components and the mixing matrix for each class are found using an extended information‐maximization algorithm, and the class membership probabilities for each pixel are computed. The pixel is allocated to the class having maximum class membership probability to produce a classification. We apply the ICAMM algorithm for unsupervised classification of images obtained from both multispectral and hyperspectral sensors. Four feature extraction techniques are considered as a preprocessing step to reduce the dimensionality of the hyperspectral data. The results demonstrate that the ICAMM algorithm significantly outperforms the conventional K‐means algorithm for land cover classification produced from both multi‐ and hyperspectral remote sensing images.