植被FAPAR的遥感模型与反演研究

FAPAR是遥感估算陆地生态系统植被净第一性生产力(NPP)的重要参数。FAPAR模型是否能真实反映植被冠层吸收光合有效辐射状况,将直接影响遥感估算植被NPP和碳循环的准确性。从FAPAR机理出发,考虑土壤反射率、冠层结构、太阳入射角等多种因素,构建了全新的定量FAPAR反演模型,并分析了太阳天顶角、LAI、土壤背景等因素与FAPAR的关系。与蒙特卡罗模拟结果的对比和用地面实测数据的验证表明该模型拥有较高的精度。选择甘肃张掖盈科灌区为研究区,利用PROBA-CHRIS高光谱多角度数据反演得到了LAI和FAPAR,并用同步观测数据验证了反演结果。     

遥感信息与作物生长模型的耦合应用研究进展

卫星遥感技术具有快速、宏观、准确、客观、及时、动态等特点,在大范围作物长势监测和产量预测等方面具有得天独厚的优势。但遥感监测常常受卫星遥感数据空间分辨率、时间分辨率等因素的影响,且遥感信息大多反映的是瞬间物理状况。作物生长模型是对作物生长、发育、产量形成过程中的一系列生理生化过程进行数学描述,是一种面向过程、机理性的动态模型。但是,当作物模拟从单点研究发展到区域应用时,由于随空间尺度的增大导致模型中一些宏观资料的获取和参数的区域化方面存在很多困难。
遥感信息与作物生长模型的耦合应用可以解决作物长势监测和产量预测等一系列农业问题,越来越受到相关研究人员的关注,已经逐渐成为一个重要的研究领域。因此,随着作物模型和遥感技术的迅速发展,如何将两者结合,进行互补性的研究是很有意义的。在查阅了相关资料的基础上,综述了遥感信息与作物生长模型的耦合应用以及发展历程,分别阐述了两种遥感数据与作物生长模型的结合方法——强迫法和同化法,总结了两类方法的应用情况。最后提出了该领域存在的问题,以及进一步解决的研究方向。     

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。     

多尺度纹理及多光谱影像协同的遥感岩性识别方法

岩性识别是遥感图像分类的难点,也是遥感地质应用的难点和热点。从遥感地质应用的实际需求出发,以西昆仑地区侏罗纪沉积岩地层为例,通过尺度转换提取高分遥感图像的多尺度纹理信息,采用波段叠加的方式协同多尺度纹理信息与ASTER影像多光谱信息进行岩性识别方法研究。利用WorldView-2全色数据进行向上尺度转换,形成空间分辨率分别为0.5,2,5,10,15,30m6种尺度图像数据,基于灰度共生矩阵提取各尺度上的纹理信息;将不同尺度的纹理信息分别与ASTER多光谱数据叠加形成协同数据;采用监督分类方法对研究区协同数据进行岩性分类。结果表明:(1)岩性纹理信息对空间尺度表现出依赖性,纹理信息量及含义随空间尺度不同而变化;(2)每套特定岩层因其独特的几何空间结构特征(厚度、产状、夹层、互层等)都有与之相适应的最佳纹理尺度,且该最佳尺度下纹理与光谱的协同效应最大;(3)纹理信息与多光谱数据形成的协同数据能有效提高岩性分类的精度,分类精度提高的程度与纹理计算的尺度相关。研究区岩性分类结果显示当纹理尺度为10m时,与仅基于ASTER纯光谱分类结果相比,精度提高了约6.9%。     

基于遗传算法优化BP神经网络的玉米遥感估产方法

以内蒙古自治区开鲁县玉米作物为研究对象,将生育期内玉米遥感影像所提取的多种植被指数和实地采样点的测产数据作为训练值,利用BP(back propagation)神经网络和遗传算法优化BP(GA-BP)神经网络估产模型,得出网络预测的玉米产量数值。通过决定系数R 2和均方根误差RMSE,比较实测产量与预测产量之间的精度,BP神经网络模型R^2为0.8452,RMSE(%)为28.37;遗传算法优化BP神经网络模型R^2为0.9850,RMSE(%)为6.70,表明遗传算法优化BP神经网络估产模型具有一定可行性和可信度。     

甘肃花牛山金矿区热红外高光谱遥感特征分析与找矿预测

利用甘肃省花牛山金矿区热红外高光谱遥感数据(CASI/SASI/TASI)提取矿化蚀变矿物进行矿物填图,通过地面热红外高光谱测试验证其准确性,对热红外高光谱遥感矿化蚀变矿物分布规律和组合特征进行分析,结合矿床矿体特征、围岩蚀变特征,建立基于CASI/SASI/TASI数据蚀变矿物分带找矿预测模型。利用该模型在找矿预测区内进行找矿预测,新发现多处金、铜等多金属矿化线索,找矿效果显著,为热红外高光谱遥感蚀变矿物填图技术的推广应用提供了实例。     

高光谱技术提取植被生化参数机理与方法研究进展

概述了目前利用高光谱技术估测地表植被生化参数理论与技术的最新研究进展,着重介绍了前人为提高遥感精度不断改进从光谱数据中提取植被生化参数的一些方法和理论,重点论述了提高遥感信息的信噪比（SNR）、改进遥感数据的分析方法、植被物理参数的细化和逐步确定,是目前植被生化参数遥感估测研究的前沿领域和科学问题,为人们尽快全面了解高光谱技术在植被生化参数方面应用进展和方法拓展,提供了条件、概貌和综论。     

基于多尺度循环注意力网络的遥感影像场景分类方法

高分辨率遥感影像场景分类一直是遥感领域的研究热点.针对遥感场景对尺度的需求具有多样性的问题,提出了一种基于多尺度循环注意力网络的遥感影像场景分类方法.首先,通过Resnet50提取遥感影像多个尺度的特征,采用注意力机制得到影像不同尺度下的关注区域,对关注区域进行裁剪和缩放并输入到网络.然后,融合原始影像不同尺度的特征及其关注区域的影像特征,输入到全连接层完成分类预测.此分类方法在UC Merced Land-Use和NWPU-RESISC45公开数据集上进行了验证,平均分类精度较基础模型Resnet50分别提升了1.89%和2.70%.结果表明,多尺度循环注意力网络可以进一步提升遥感影像场景分类的精度.      

马达加斯加索菲亚省BINARA地区铬铁矿遥感找矿预测

马达加斯加北部地区地广人稀、交通不便、植被稀疏、矿产资源丰富,利用遥感技术提取矿化蚀变信息具有独特的优势和较好的效果。为了对BINARA地区铬铁矿进行遥感找矿预测,文章首先利用Landsat 8遥感数据2、4、5、6波段进行主成分分析提取研究区内基性-超基性岩信息;然后基于实测高光谱数据,采用偏最小二乘回归法建立Fe含量定量反演模型,并利用该模型对Hyperion高光谱遥感数据进行计算,提取研究区内Fe含量异常信息;最后根据综合遥感异常信息以及野外调查验证结果,确定研究区内的Fe含量为一级异常的区域作为预测区。     

遥感植物地球化学方法的研究进展

近些年来随着遥感技术的不断发展,在植被覆盖区的金属矿床的探查中,遥感植物地球化学方法得到了越来越广泛的应用。文章从理论、技术和应用三个方面对遥感植物地球化学方法的发展进行阐述,并指出目前存在的问题。其中使用定量化的手段提取植物地球化学信息是未来的必由之路。然而在定量化的过程中,光谱尺度效应和空间尺度效应是两个重要的影响模型精度的因素。针对这个问题,文章提出两种解决方案来减小光谱尺度效应的影响,然而这两种方法的可操作性以及可靠性还需要进一步探讨。最后,文章对遥感植物地球化学方法的发展前景进行阐述,并指出在以后定量模型的建立中可以引入支持向量机、投影追踪等机器学习的方法。     

Spectral wheat yield prediction modeling using SPOT satellite imagery and leaf area index

As wheat represents the main staple food and strategic crop in Egypt and worldwide and since remote sensing satellite imagery is the tool to obtain synoptic, multi-temporal, dynamic, and time-efficient information about any target on the Earth, the main objective of the current study is to use remote sensing satellite imagery to generate remotely sensed empirical preharvest wheat yield prediction models. The main input parameters of these models are spectral data either in the form of spectral reflectance data released from Satellite Pour l’Observation de la Terre (SPOT) 4 satellite imagery or in the form of spectral vegetation indices. The other input factor is leaf area index (LAI) that was measured by LAI Plant Canopy Analyzer. The four spectral bands of SPOT4 imagery are green, red, near-infrared, and middle infrared; the five vegetation indices that are forms of ratios between red and near-infrared bands are normalized difference vegetation index, ratio vegetation index, soil-adjusted vegetation index, difference vegetation index, and infrared percentage vegetation index. Another vegetation index is green vegetation index that is calculated through a ratio between green band and near-infrared band. Each of the above-mentioned factors was used as an input factor against wheat yield to generate wheat yield prediction models. All generated models are site-specific limited to the area and the environment and could be applicable under similar conditions in Egypt. The study was carried out in Sakha experimental station by using the dataset from two wheat season 2007/2008 and 2009/2010. The total wheat area was 1.3 ha cultivated by Sakha 93 cultivar. Modeling and validation process were carried out for each season independently. Modeled yield was tested against reported yield through two common statistical tests; the standard error of estimate between modeled yield and reported yield, and the correlation coefficient for a direct regression analysis between modeled and reported yield with each generated model. Generally, as shown from the correlation coefficient of the generated models, green and middle infrared bands did not show good accuracy to predict wheat yield, while the other spectral bands (red and near-infrared) bands showed high accuracy and sufficiency to predict yield. This was proven through the correlation coefficient of the generated models and through the generated models with the wheat crops for the two seasons. Accordingly, the green vegetation index that is generally calculated from green and near-infrared bands showed relatively lower accuracy than the rest of the vegetation index models that are calculated from red and near-infrared bands. LAI showed high accuracy to predict yield as shown from the statistical analysis. The models are applicable after 90 days from sowing stage and applicable in similar regions with the same conditions.     

Image based remote sensing method for modeling black-eyed beans (Vigna unguiculata) Leaf Area Index (LAI) and Crop Height (CH) over Cyprus

In this paper, Leaf Area Index (LAI) and Crop Height (CH) are modeled to the most known spectral vegetation index — NDVI — using remotely sensed data. This approach has advantages compared to the classic approaches based on a theoretical background. A GER-1500 field spectro-radiometer was used in this study in order to retrieve the necessary spectrum data for estimating a spectral vegetation index (NDVI), for establishing a semiempirical relationship between black-eyed beans’ canopy factors and remotely sensed data. Such semi-empirical models can be used then for agricultural and environmental studies. A field campaign was undertaken with measurements of LAI and CH using the Sun-Scan canopy analyzer, acquired simultaneously with the spectroradiometric (GER1500) measurements between May and June of 2010. Field spectroscopy and remotely sensed imagery have been combined and used in order to retrieve and validate the results of this study. The results showed that there are strong statistical relationships between LAI or CH and NDVI which can be used for modeling crop canopy factors (LAI, CH) to remotely sensed data. The model for each case was verified by the factor of determination. Specifically, these models assist to avoid direct measurements of the LAI and CH for all the dates for which satellite images are available and support future users or future studies regarding crop canopy parameters.     

考虑水分胁迫后效应的作物水模型

水分胁迫具有后效应,前期水分胁迫可以影响作物后期叶面积和需水量的增减。以Jensen作物 水模型为基础,引入了水分胁迫后效应影响系数,对原模型进行了修正。修正后的模型可以将阶段水分胁迫与前期胁迫后效应对产量的影响加以区分,避免了原模型中可能产生的虚缺水现象,并可对作物(以玉米为例)前期水分胁迫处理后,后期需水量增加以及苗期胁迫处理可维持较高产量的原因进行合理解释。通过田间试验结果分析,改进后模型的模拟结果符合实际,并具有较好的精度。对模型存在的问题和不足也进行了探讨。     

作物水分生产函数与农田非充分灌溉研究述评

介绍作物水分生产函数国内外研究现状,把目前通用的各类模型归纳为最终产量模型和动态产量模型两大类,分析评述两类模型的特点及适用条件.此外,对作物水分生产函数的试验方法和试验处理设计作了介绍.     

气候模式中关键陆面植被参量遥感估算的研究进展简

Satellite remote sensing data play an important role in the improvement of climate models forcing field, relevant physical parameters and simulation accuracy. At present, there are many years of satellite remote sensing data and a variety of products about land surface attributes. However, the application of satellite remote sensing data to climate models is still very limited. Fully using satellite remote sensing data is important to improving the simulation ability. In the paper, remote sensing estimates methods of three key land surface parameters including Fractional Vegetation Coverage（FVC）, Leaf Area Index（LAI）and surface albedo（Albedo）is reviewed and up or down scaling land surface variables in validation process is analyzed. Secondly, taking WRF（Weather Research and Forecasting）model as an example, three parameters in climate model are described. Finally, the key problems of using remote sensing data in climate models are discussed, which comprise the uncertainties and scales of remote sensing estimation parameters and the future direction is prospected.     

Importance of soil and elevation characteristics for modeling hurricane-induced power outages

Hurricanes can severely damage the electric power system, and therefore, predicting the potential impact of an approaching hurricane is of importance for facilitating planning and storm-response activities. A data mining approach, classification and regression trees (CART), was employed to evaluate whether the inclusion of soil and topographic variables improved the predictive accuracy of the power outage models. A total of 37 soil variables and 20 topographic variables were evaluated in addition to hurricane, power system, and environmental variables. Hurricane variables, specifically the maximum wind gust and duration of strong winds, were the most important variables for predicting power outages in all models. Although the inclusion of soil and topographic variables did not significantly improve the overall accuracy of outage predictions, soil type and soil texture are useful predictors of hurricane-related power outages. Both of these variables provide information about the soil stability which, in turn, influences the likelihood of poles remaining upright and trees being uprooted. CART was also used to evaluate whether environmental variables can be used instead of power system variables. Our results demonstrated that certain land cover variables (e.g., LC21, LC22, and LC23) are reasonable proxies for the power system and can be used in a CART model, with only a minor decrease in predictive accuracy, when detailed information about the power system is not available. Therefore, CART-based power outage models can be developed in regions where detailed information on the power system is not available.     

Atmospheric pressure patterns,climatic change and winter wheat yields in North America

Winter wheat yields over a large area of the United States Great Plains are described as functions of the monthly surface atmospheric pressure pattern over North America. Seventy-eight years of pressure data were spatially decomposed with principal components analysis, and linear combinations of the resulting eigenvectors were used to fit a time series of five sensible climatic variables that are generally considered to be important to wheat yields. This surrogate data set was then used to fit the yield time series. The yield data were initially detrended with a simple linear estimator and adjusted with base constants specific for each crop reporting district.Interannual variation in the pressure field explains approximately 40% of the remaining variance in the yield data over the states of Colorado, Kansas, Oklahoma, Nebraska and Texas. When a reduced model was tested on five years of data simultaneously withheld, a similar amount of the variance was explained. Twentyone eigenvectors are consistently associated with the twelve sensible climatic parameters to which yields are most sensitive. Of these, five were found to be significantly changing (at between the 95 and 99% levels), in either linear or quadratic fashion, over the length of the pressure record. It is concluded that long-term change can and does take place in features of the general circulation that are important determinants of large area crop yields.     

基于HJ-CCD影像数据的新疆喀纳斯自然保护区植被叶面积指数估算

以新疆喀纳斯自然保护区为研究区, 评价了HJ-CCD影像数据估算植被叶面积指数(LAI)的能力及其对大气订正方法的敏感性.分别利用6S和FLAASH两种大气订正模型对HJ1B-CCD2影像进行大气订正, 比较了大气订正前后不同植被(针叶林、阔叶林、针阔混交林和草地)反射率及5种植被指数(NDVI、SR、SAVI、MSR、ARVI)的变化, 进而建立了4种植被类型LAI的遥感估算模型, 分析了LAI的空间分布格局.结果表明: 大气订正后可见光波段的反射率降低, 6S模型订正后近红外波段的反射率上升, 而FLAASH模型订正后近红外波段的反射率下降.大气订正后NDVI、SR、SAVI(除针叶林)和MSR上升, 6S模型订正后所有植被类型的ARVI下降, FLAASH模型订正后针叶林和阔叶林的ARVI上升, 而针阔混交林和草地的ARVI下降.大气订正提高了植被指数与LAI之间的相关性, 对于针叶林、阔叶林、针阔混交林而言, 利用6S模型订正后的反射率建立的模型优于FLAASH模型订正后的反射率建立的模型, 而草地却相反.经过大气订正, HJ-CCD影像数据可应用于研究区植被LAI的估算.研究区LAI的高值集中在湖泊和河流附近, 低值分布在海拔较高处.山地森林草原带、亚高山森林带、高山灌丛草甸带、高山冻原、高山冰川带植被LAI的平均值分别为2.6、3.9、2.5、1.7和1.0.     

冬小麦水肥生产函数的Jensen模型和人工神经网络模型及其应用

作物水肥生产函数研究是非充分灌溉理论的重要内容,也是提高农田水、肥利用效率的基础.在作物水分生产函数Jensen模型的基础上,引入肥料因子构造了水肥生产函数的Jensen模型;同时构造了作物水肥生产函数的人工神经网络模型.利用北京地区冬小麦田间试验资料对以上2个模型进行了分析,表明以上模型均可用于描述水分、肥料等因素对作物产量的影响,进而可对作物产量进行预测,且模型都具备一定的精度.