Correlation －SVM 模型在大坝变形预测中的应用

引起大坝变形的影响因素很多,即在利用支持向量机（ SVM）模型进行大坝变形分析和预报的过程中,需要将所有的影响因子都输入到SVM模型中,这样会造成输入因子的不侧重性,基于此,本文对大坝变形的影响因子进行相关性分析,根据大坝变形影响因子和大坝变形量之间的关系来确定最优的影响因子,即将比重比较大的影响因子输入到SVM模型中,从而提高了SVM模型运行效率及预测的精度和速度。     

一种基于主成分分析的时空地理加权回归方法

针对时空地理加权回归模型(GTWR)进行预测时,输入变量较多导致计算复杂度高,而输入变量较少引起预测精度降低这一问题,提出了一种基于主成分分析的时空地理加权回归方法(PCA-GTWR)。该方法采用非线性主成分分析方法,先对影响PM2.5浓度的若干相关变量降维处理得到几个综合指标,并将其作为GTWR模型的输入变量进行预测。为验证该方法的有效性,采用北京市2014-04—2017-03的PM2.5数据,利用Pearson相关系数法选取与PM2.5浓度具有较高相关性的影响因素作为常规的GTWR模型的输入变量,在变量个数相同的前提下,与本文方法进行对比。结果表明应用非线性主成分分析方法对相关变量进行预处理后,有效地解决了变量之间的共线性,保留了原始影响因素主要信息,提高了运算效率,且该方法的MAE、RMSE、AIC均低于常规的GTWR模型,拟合优度GF最高达到88.11%。     

PCA和布谷鸟算法优化SVM的遥感矿化蚀变信息提取

为了进一步提高遥感矿化蚀变信息提取的精度,本文提出了一种基于主成分分析PCA (Principal Component Analysis)和布谷鸟算法优化支持向量机SVM (Support Vector Machine)的遥感矿化蚀变信息提取方法。首先,通过波段比值法增强研究区遥感图像中的矿化蚀变信息,并获得比值图像;然后,对比值图像进行主成分分析,进而提取训练样本;接着,利用SVM对训练样本进行训练,同时采用布谷鸟算法求取SVM的最优核参数及惩罚因子,构造最优SVM模型;最后,运用最优SVM模型完成矿化蚀变信息提取。选择青海省五龙沟地区为研究区,提取羟基及铁染蚀变信息。实验结果表明,与主成分分析法、基于光谱角法和SVM的方法、基于粒子群和SVM的方法及基于波段比值、PCA和粒子群优化SVM的方法等4种方法相比,本文方法获得的遥感矿化蚀变信息和已知矿点的吻合度最高,提取效果最好。     

特征维数对支持向量机分类器性能影响的研究——以高光谱遥感影像为例

本文为验证SVM对高维特征的适应性和可靠性,针对不同特征提取方法与特征组合,以国产OMISⅡ传感器获得的北京昌平地区高光谱遥感据为例,对SVM分类器中特征维数对分类准确率的影响进行了试验,通过对主成分分析、最小噪声分离算法、相关系数分组后特征提取、导数光谱等的分析,表明SVM分类器的分类精度随着特征维数波动,其中主成分分析降维后提取的特征具有用于分类能够获得最高的准确率。通过与最大似然法和光谱角制图分类算法的比较,说明在同样的特征输入情况下SVM分类算法分类的准确率高于最大似然法和光谱角制图分类器。     

基于决策树的CBERS遥感影像分类及分析评价

以江苏省徐州市为研究区,以城市土地利用遥感分类为目标,采用CBERS多光谱数据的近红外波段、全球环境监测植被指数(GEMI)、归一化植被指数(NDVI)及主成分分析得出的第一和第二主成分作为分类的特征数据,基于先验知识和统计分析构建层次分类决策树,进而发展和改进了决策树交互式构建算法,实现了城市土地利用遥感分类。通过与最大似然分类器(MLC)和支持向量机分类器(SVM)分类结果的比较分析,表明基于多种特征的决策树分类器能够有效应用于CBERS遥感数据分类,在研究区具有良好的推广性。     

结合PCA、多尺度分割及SVM的ASTER遥感蚀变信息提取

为了利用遥感影像进行更加精确的找矿预测,本文选择新疆东天山尾亚地区ASTER数据进行矿化蚀变信息提取方法研究。为了提高信息提取精度,本文提出了结合主成分分析(PCA)、多尺度分割和支持向量机(SVM)的遥感矿化蚀变信息提取方法。首先,分析ASTER数据的特征,选取各矿化蚀变信息的特征波段,对组合波段进行主成分分析,获得主分量图像;然后,对各主分量图像进行多尺度分割,并获得分割之后的均值图像;接着,提取训练样本,利用SVM对训练样本进行训练,采用试验方法求得最优核参数和松弛变量,构造最优SVM模型;最后,运用最优SVM模型完成矿化蚀变信息的提取。进行主成分分析时,铁染蚀变信息选择Band1、2、3、4组合,Al-OH基团蚀变信息选择Band 1、4、6、7组合,OH和CO32-基团蚀变信息采用Band 1、2、8、9组合。在运行SVM时采用了序列最小优化算法(SMO)进行求解,速度提高了12%。实验结果表明,与波段比值法、主成分分析法及基于光谱角和SVM的方法等3种方法相比,本文方法提取铁染蚀变信息、Al-OH基团蚀变信息及OH和CO32-基团蚀变信息的总体精度可达到87.98%、 90.01%及88.93%,Kappa系数分别为0.8011、0.8134及0.8023,与成矿区带、已知矿点和已有不同地质背景成矿特征相关性较好。     

基于主成分分析的GA-BP神经网络地表下沉系数预测

为提高条带开采地表下沉系数预测准确率,基于地表下沉系数影响因素具有一定相关性、不确定性以及非线性的复杂现象,建立基于主成分分析的遗传算法(Genetic Algorithm,GA)优化的神经网络GA-BP智能预测模型.利用遗传算法对BP神经网络(Back Propagation Neural Network)的初始权值和阈值进行优化处理,通过SPSS20(Statistical Product and Service Solutions 20)软件对地表下沉系数影响因素进行主成分分析,降低数据维度,消除变量间的冗余信息,找出主成分并作为模型的输入样本,利用MATLAB(Matrix Laboratory)软件进行仿真与分析.结果表明:与传统BP神经网络模型和主成分分析(Principal Component Analysis,PCA)的PCA-BP神经网络模型相比,基于主成分分析的GA-BP模型的相对误差不超过5％,与实测值更为接近,预测精度进一步提高,基本满足矿区实际工程需要,为条带开采地表下沉系数预测提供了又一种准确可行的方法.     

基于改进SVM算法的高分辨率遥感影像分类

针对面向对象高分辨率遥感影像分类样本维数多、数据量大的特点,提出了一种简单的支持向量机(support vector machine,SVM)改进算法。首先对原始样本数据进行主成分分析(principal component analysis,PCA)实现降维,对降维后的样本数据进行SVM分类器训练,利用网格搜索法得出降维数据的最佳参数;以此参数作为基准,对基于原始样本数据的SVM分类器参数搜索范围进行重新设定,从而快速获取原始样本数据的最佳SVM分类器参数,并实现分类。利用2景World View2高分辨率影像分别对城市土地利用以及林木树种进行分类实验,比较分析传统SVM算法、仅基于PCA降维样本数据的SVM算法以及改进的SVM算法在分类精度与效率方面的差异。实验结果表明,改进的SVM算法能够快速有效地寻找最佳SVM分类器参数,并获得较高的分类精度。     

北京市热环境时空分异与区划

城市热环境空间区划是采用分区管理的思路来缓解城市社会经济发展与热环境之间矛盾的技术基础。本文构建城市热环境区划模型的思路为:(1)将不同时相的MODIS地表温度数据产品进行正规化、分级,分析2008年北京城市热环境时空分布特征。(2)构建城市热环境影响因素评价体系,并通过空间主成分分析计算得到热环境影响主成分因子。(3)通过自组织映射神经网络,利用热环境影响主因子,进一步对热环境进行空间区划。结果表明,北京夜间较白天城市热岛分布层次感明显,夏季白天较其他季节高温区聚合程度高。区域下垫面组成要素直接影响热环境,北京城市热环境的主成分因子依次为植被覆盖、地形地貌、城市下垫面建设规模和人为热排放,并依此将北京划为7个热环境区域,根据各个分区热环境成因机制差异分别提出热环境改善和调控措施。     

主成分分析模型在数据处理中的应用

主成分分析模型是一种将原始多个指标转化为少数几个相互独立的包含原始指标绝大部分信息的综合指标的统计学方法。利用MATLAB软件探索了利用主成分分析模型对专题数据进行处理的方法,并利用科学可视化方法对处理结果进行分析评价。实验表明,主成分分析模型能够有效简化原始变量,挖掘原始数据中的隐藏信息。     

Enhancing the Reliability of 3D Subsurface Models through Differential Weighting and Mathematical Recombination of Variable Quality Data

One of the first stages of the three‐dimensional (3D) subsurface modeling process involves collation and analysis of available borehole and/or outcrop data to identify individual subsurface units, usually distinguished by the grain size of the sediment, and the elevation of their bounding contacts. Input data can come from a variety of sources and may be categorized according to their reliability and/or quality. The output from the 3D model is a prediction of subsurface conditions based on these data and the reliability of the output model is highly dependent on both the quality of input data and the types of interpolation methods used. This article presents a new quality weighting methodology that allows the user to assign a differential weighting factor to data points of variable quality in the modeling process. Input data are categorized into high and low quality datasets which are then recombined using a grid math process in which a differential “weighting” factor is applied. This allows the 3D modeling program to maximize the use and effectiveness of data from all available sources while giving high quality data greater influence on the final model output, and will result in the generation of more accurate and reliable 3D subsurface models.     

Derivation of tree skeletons and error assessment using LiDAR point cloud data of varying quality

The architecture of trees is of particular interest for 3D model creation in forestry and ecolocical applications. Terrestrial (TLS) and mobile laser scanning (MLS) systems are used to acquire detailed geometrical data of trees. Since 3D point clouds from laser scanning consist of large data amounts representing uninterpreted topographical information including noise and data gaps, an extraction of salient tree structures is important for further applications. We present a fully automated modular workflow for topological reliable reconstruction of tree architecture. Object-based point cloud processing such as branch extraction is combined with tree skeletonization. Branch extraction is performed using a segmentation procedure followed by segment-based analysis of form indices derived from eigenvector metrics. Extracted branch primitives are simplified and connected to line features during skeletonization. The modular workflow allows comprehensive parameter tests and error assessments that are used for a calibration of the module parameters with respect to various characteristics of the input data (e.g noise, scanning resolution, and the number of scan positions). The estimated parameter settings are validated using an exemplary MLS data set. The quality of input point cloud data, strongly influencing the quality of the skeleton results, can be improved by the presented branch extraction procedure. The potential for data improvement increases with increasing point densities. For our object-based appoach, we can show that the presence of erroneous structures and filtering artifacts have the strongest influence onto the quality of the derived skeletons. In contrast to traditional skeletonization approaches, the existance of data gaps has less influence onto the results.     

基于GIS的城市空气污染综合分析

在城市空气污染的模型研究中,GIS可以作为复杂环境模型的输入数据处理器,而且GIS有一个重要的功能模块——统计分析模块,GIS能够对模型的运算输出数据进行综合处理,并对结果进一步分析,进行有关的查询以及对研究成果的可视化输出。在基于GIS的城市空气污染的综合分析中,我们探究了城市土地利用现状对空气质量的影响,本文采用模糊综合分析法探索城市三大污染源在空气污染中所占的比重,制作城市空气污染等级分布图。GIS在大气污染研究领域可以充分发挥其空间分析和可视化优势,为治理工作提供技术路线。     

光学遥感影像表观辐亮度地形效应纠正物理模型

针对已有地形纠正方法的不足, 在山区辐射传输模型简化的基础上, 提出了水平地面上接收到的漫射辐射与垂直于太阳方向表面接收的直射辐射比例因子的概念, 建立了仅需要太阳角度信息和大气模式作为输入参数, 主要针对地形效应本身进行纠正的简单纠正模型, 可以将复杂地形区光学遥感影像表观辐亮度纠正为无地形影响的水平地表辐亮度, 并以TM影像为例进行了实验验证。     

水文气象因素对中国北疆流动重力变化的影响

水文气象因素引起的重力变化是影响地震重力变化成果解释的重要因素。以中国北疆地区为研究区域,借助全球陆地数据同化系统（global land data assimilation systems,GLDAS）全球水文模型数据、大气模型数据,计算2016-01—2017-12时段内水文气象因素对研究区域的重力影响。计算结果表明,陆地水影响的年变化为1.3 μGal,两期陆地水影响空间分布的差异低于1 μGal;大气影响的年变化为8 μGal,两期大气影响空间分布的差异达到6 μGal。利用2016-04、2016-08和2017-06三期流动重力测量数据,对比扣除水文气象因素前后的重力变化,可以看出,在中国北疆流动重力数据处理中,大尺度水文因素可以不予考虑,气象因素应予考虑。同时,为更好分析流动重力变化,建议流动重力测量过程中同时开展测点附近的土壤湿度、大气气压等观测。     

AnnGIS: Integration of GIS and a Continuous Simulation Model for Non-Point Source Pollution Assessment

Non-Point Source (NPS) models and monitoring data are often used to evaluate management practices and develop NPS pollution control plans. Application of a dynamic NPS model requires efficient input data acquisition, storage, organization, reduction, and analysis accompanied by manipulation, interpretation, reporting, and display of model outputs. A Geographic Information System (GIS) helps extract, store, and organize input data as well as manipulate and display model outputs. This paper illustrates the development of an integrated GIS system for a continuous simulation, pollutant-loading model, AnnAGNPS ( Ann ualized AG ricultural N on- P oint S ource Pollution). The integrated system, called AnnGIS, was developed using the ArcView GIS and related program extensions. Using AnnGIS, modeling studies and management plans can be efficiently and easily developed. AnnGIS helps store, organize, and manipulate spatial and tabular data, extract spatial input parameters, develop analysis scenarios, and visualize input and output data in spatial, tabular, and graphical forms. AnnGIS is generic in nature (not limited to a particular geographic location) and can be successfully used in regions for which AnnAGNPS is designed. AnnGIS's powerful graphical user interface and reference data sets facilitate efficient and informed decision-making concerning agricultural non-point pollution control and management.     

Effect of source-varying input data on erosion potential model performance

This paper examines the performance of the Erosion Potential Method (EPM or Gavrilovi?) and the model response to input data variations caused by choosing different sources of information for the same parameter. The research presented addresses the input data uncertainty via an analysis of the two model input parameters (the soil protection coefficient and the soil erodibility coefficient). The parameter uncertainty analysis is performed following two different approaches: uncertainty analyses of both the selected sample size and the entire population are conducted for an erosion assessment case study of the Dubra?ina River catchment, Croatia. The analysis indicated that, when changing the data source, significant changes in the model outcome values can occur (up to 47% for this case study). Future method modifications should consider the mitigation of these two parameters by potentially making structural changes in the model and therefore moderating the effect.     

基于参数不确定性分析的SAR土壤水分反演精度控制方法

参数不确定性是SAR反演土壤水分的重要不确定性来源,为控制土壤水分反演精度,提出一种基于参数不确定性的有效控制土壤水分反演精度的方法,使用该方法可以控制参数的误差范围。首先使用全局敏感性分析方法,确定后向影响散射系数输出的主要参数;在不同量级高斯噪声随机扰动下,将大量各参数采值输入AIEM模型中,得到带噪声的后向散射系数集合;再使用LUT法反演土壤水分,计算反演结果满足误差量级控制范围。以此为基础,利用ENVISAT ASAR双极化数据(VV、VH)和实测土壤水分数据进行验证,利用LUT法反演得到带噪声的土壤水分,计算ASAR影像中采样点土壤水分反演值RMSE0.04cm3/cm3。结果表明各影响参数误差量级控制范围可有效控制土壤水分反演精度,在较大的入射角范围内都适用。     

Digital Terrain Model Resolution and its Influence on Estimating the Extent of Rockfall Areas

As rockfall can cause a great deal of damage, it is essential to know its spatial propagation. Rockfall models are sensitive to the resolution of input data, i.e. the Digital Terrain Model (DTM) used. Nowadays, high resolution elevation data are available area‐wide from airborne laser scanning (ALS). However, rockfall models are designed for analysis on a certain scale, which means that high resolution input might not necessarily improve model results (e.g. for regional scale studies). Our aim is to estimate the reach of rockfall by analysing different input resolutions of an ALS DTM. The presented empirically–based model, implemented in Python 2.7, is a modified version of the zenital method including an iterative random walk trajectory model, which is designed for rockfall hazard assessment at the regional scale. Trajectories and rockfall probability maps are generated for selected DTM input resolutions. The comparison shows that high resolution DTMs do consider local topography better and thus lead to more realistic results than low resolution DTMs.     

地球自转参数的RFFLS短期预报算法研究

地球自转参数(ERP)是实现地心天球坐标系(geocentric celestial reference system,GCRS)与国际地球坐标系(international terrestrial reference system,ITRS)相互转换的必要参数,是国际GNSS服务组织(IGS)和国际GNSS监测评估系统(iGMAS)分析中心的重要产品。本文针对最小二乘地球自转参数预测算法会造成数据饱和以及新旧数据在数据处理及预报中被同等对待等问题,将遗忘因子引入最小二乘预测算法,进而提高ERP预报精度。遗忘因子递推最小二乘算法能防止数据饱和,降低旧数据的影响,加强新数据的作用,降低在求解拟合参数时出现秩亏矩阵求逆的几率,提高预报精度。本文详细推导了遗忘因子递推最小二乘表达式,探究了最佳遗忘因子,并通过ERP试验将该方法和原最小二乘的试验结果及LS-AR模型的预报结果作对比,发现仅用遗忘因子最小二乘模型预测就可以达到与LS-AR组合模型预测相当的精度。