摘要
为了及时、准确的了解清河水库总悬浮物浓度情况,采用Landsat卫星OLI数据,通过SPSS软件分析计算OLI数据的单波段及波段组合与总悬浮物浓度之间的相关关系,选取相关系数最大者分别构建比值线性回归模型和非线性的最小二乘支持向量机模型(LS-SVM),对清河水库总悬浮物浓度进行了遥感定量反演研究。结果表明,相比于比值线性回归模型,LS-SVM模型将预测值与实际值的可决系数R2从0.686提高到0.88,平均相对误差从3.52%减小到3.16%,利用LS-SVM模型对总悬浮物浓度的反演精度显著提高。
To timely and accurately monitor Qinghe Reservoir total suspended matter(TSM),with the OLI data of Landsat satellite and the calculation of SPSS,this paper analyzes the relevant relationship between TSMconcentration and single band or band combinations,the ratio of the linear regression model and nonlinear least squares support vector machine(LS-SVM) is established by taking the relation as the correlation coefficient of largest,then quantitative remote sensing inversion research is done on TSMconcentration of Qinghe Reservoir. Results indicate that the LS-SVMmodel can make predictions and actual value of determination coefficient R2 increase from 0.686 to 0.88 and the average relative error decreases from 3.52% to 3. 16% compared with the ratio linear regression model. Therefore,the LS-SVMmodel is used for the inversion precision of TSMincrease.
引文
[1]张运林,秦伯强,陈伟民,等.太湖水体中悬浮物研究[J].长江流域资源与环境,2004,13(3):266-271.
[2]查桂红.基于GOCI影像的内陆水体悬浮物浓度遥感估算研究[D].南京:南京师范大学,2013.
[3]Ostlund C,Flink P,Strombeck N,et al.Mapping of the water quality of Lake Erken,Sweden,from Imaging Spectrometry and Landsat Thematic Mapper[J].The Science of the Total Environment,2001,268:139-154.
[4]马驰,姜琦刚,李远华,等.利用ASTER影像反演松辽平原水体悬浮物浓度[J].水资源保护,2009,25(2):72-74.
[5]Binding C E,Bowers D G,Mitchelson-Jacob E G Estimating suspended sediment concentration from ocean colour measurements in moderately turbid waters;the impact of variable particle scattering properties[J].Remote Sensing of Environment,2005,94:373-383.
[6]宋庆君,马荣华,唐军武,等.秋季太湖悬浮物高光谱估算模型[J].湖泊科学,2008,20(2):196-202.
[7]Dekker A G.Analytical algorithms for lake water TSM estimation for retrospective analyses of TM and SPOT sensor data[J].Journal of Remote Sensing,2002,23:15-35.
[8]Miller R L,Mc Kee B A.Using MODIS Terra 250 m imagery to map concentration of total suspended matter in coast waters[J].Remote Sensing of Environment,2004,93(1-1):259-266.
[9]Dekker A G,Vos R J,Peters S W M.Comparsion of remote sensing data,model results and in situ data for total suspended matter(TSM)in the southern Frisian lakes[J].Science of the Total Environment,2001,268(1-3):197-214.
[10]温小乐,徐涵秋.基于多元同步数据的闽江下游悬浮物定量遥感[J].环境科学,2008,29(9):2 441-2 447.
[11]管义国,王心源,吉文帅.巢湖水体悬浮物的遥感分析[J].遥感信息,2007,(5):39-43.
[12]张毅博,张运林,査勇,等.基于Landsat8影像估算新安江水库总悬浮物浓度[J].环境科学,2015,36(1):56-63.
[13]李渊,李云梅,施坤,等.基于光谱分类的总悬浮物浓度估算[J].光谱学与光谱分析,2013,33(10):2 721-2 726.
[14]李颖.基于环境1号卫星多光谱数据的太湖总悬浮物浓度估算模型研究[D].南京:南京师范大学,2011.
[15]刘大召,陈楚群,刘汾汾,等.利用混合光谱分解估测珠江口悬浮泥沙浓度[J].热带海洋学报,2009,28(5):43-48.
[16]Williams P C.Implementation of near-infrared technology:in the agricultural and food industries[C]∥USA:American Association of Cereal Chemists,Ins,2001.
[17]徐红敏.基于支持向量机理论的水环境质量预测与评价方法研究[D].长春:吉林大学,2007.
[18]送召青,崔和,胡云安.支持向量机理论的研究与进展[J].海军航空工程学院学报,2008,23(2):143-148,152.
[19]阎威武,邵惠鹤.支持向量机和最小二乘支持向量机的比较及应用研究[J].控制与决策,2003,18(3):358-360.
[20]李云梅,黄家柱,陆皖宁,等.基于分析模型的太湖悬浮物浓度遥感监测[J].海洋与湖沼,2006,37(2):171-177.
[21]赵碧云,贺彬,朱云燕,等.滇池水体中总悬浮物含量的遥感定量模型[J].环境科学与技术,2001,94(2):16-18.