首页 | 官方网站   微博 | 高级检索  
     

削弱水分影响的潮滩表层沉积物遥感分类方法研究
引用本文:李欢,张东,张鹰.削弱水分影响的潮滩表层沉积物遥感分类方法研究[J].海洋学报,2012,34(6):84-93.
作者姓名:李欢  张东  张鹰
作者单位:1.南京师范大学 地理科学学院, 江苏 南京 210023;奥克兰大学 地理、地质和环境科学学院 新西兰 奥克兰
基金项目:国家自然科学基金(40606044);教育部博士点基金(200803190007);江苏高校优势学科建设工程资助项目。
摘    要:潮滩沉积物中的水分掩盖了沉积物本身的光谱信息,不利于沉积物遥感分类。利用江苏大丰潮滩沉积物在不同含水量下的398条实测光谱曲线数据,基于线性光谱分解方法定量化分析了沉积物光谱对含水量变化的响应,建立削弱水分影响的潮滩表层沉积物组分(砂、粉砂、黏土)含量反演模型,应用于Hyperion高光谱遥感影像,获得沉积物组分的空间分布,并结合Shepard三角分类法实现了潮滩表层沉积物的自动分类。研究结果表明:(1)当沉积物含水量增加时,沉积物光谱中水分权重以含水量的2倍速率增加,含水量大于25%,则水分对光谱的影响占主导作用;(2)2 143 nm波段反射率对沉积物的含水量变化敏感,利用2 143 nm波段反射率进行潮滩沉积物的含水量反演,模型拟合度r2可达0.81;(3)983,1 134 nm波段反射率对沉积物组分含量反演敏感,将含水量与沉积物组分敏感波段一起建立多元线性回归方程,可有效削弱水分的影响,适用于高含水量潮滩区的沉积物组分含量反演;(4)将组分反演结果结合Shepard分类,可实现潮滩沉积物分类,得到潮滩沉积物类型的空间分布特征,分类精度为75.93%,Kappa系数为0.6。

关 键 词:沉积物分类    潮滩    高光谱遥感    含水量
收稿时间:9/2/2011 12:00:00 AM
修稿时间:2012/3/20 0:00:00

The study on a surface sediment classification methocl by weakening moisture influence in intertidal zone using remote sensing
LI Huan,ZHANG Dong and ZHANG Ying.The study on a surface sediment classification methocl by weakening moisture influence in intertidal zone using remote sensing[J].Acta Oceanologica Sinica (in Chinese),2012,34(6):84-93.
Authors:LI Huan  ZHANG Dong and ZHANG Ying
Affiliation:1.School of Geographic Science, Nanjing Normal University, Nanjing 210046, China;School of Geography, Geology and Environmental Science, University of Auckland, Auckland 1010, New Zealand2.School of Geographic Science, Nanjing Normal University, Nanjing 210046, China
Abstract:The spectral signal of sediment on intertidal flats is easily concealed as the result of high moisture inside,which is the main obstacle for retrieving sediment types by means of remote sensing technology.To improve the accuracy of quantifying sediment characteristics,the effects of varying moisture must be removed.With the use of in situ samples of sediment collected in Dafeng,Jiangsu Province,China and its spectra measured under different moisture conditions in the laboratory,a linear spectral unmixing technology is employed to obtain moisture abundance and analyze the spectral response of sediment to varied moisture content.The moisture content layer of study area retrieved by a linear regression techniology on Hyperion image is introduced as variable into multiple variable linear regression equations.The models,which can eliminate the moisture impact,are constructed and applied in Hyperion image to retrieve components'(sand,silt and clay) spatial distribution.Retrieved components are input to Shepard classification system to map the sediment types of intertidal flats automatically.The results are shown as below:(1)the spectral contribution from water is dominant while the moisture content is higher than 25% and its value increases two times with the rate of moisture content increasing;(2) the reflectance of samples in 2 143 nm is sensitive to the variation of moisture,and square (r2) of correlation coefficient of constructed moisture model is 0.81;(3)the reflectance of samples in 983 and 1 134 nm is sensitive to the content of each component,these two bands combined moisture sensitive band are utilized to construct the multiple variable linear regression equation,which can effectively remove the moisture impact on wet samples;(4)the overall classification accuracy is 75.93%,and Kappa coefficient reaches 0.60.This method is an effective way to map the spatial distribution characteristic of sediment types accurately on muddy intertidal flats.
Keywords:sediment classification  intertidal flats  hyperspectral remote sensing  moisture content
本文献已被 CNKI 等数据库收录!
点击此处可从《海洋学报》浏览原始摘要信息
点击此处可从《海洋学报》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号