物理模型结合SVM的格陵兰岛冰盖冻融探测

1.河南工业大学 信息科学与工程学院,河南 郑州 450001; 2.中国科学院 遥感与数字地球研究所,北京 100094

格陵兰岛; 冰盖冻融探测; 微波辐射计; 物理模型

Ice-sheet freeze-thaw detection based on physical model combined with SVM in Greenland
WANG Xing-dong1,2,DUAN Zhi-yong1,WANG Cheng2,LI Xin-guang1

(1.College of Information Science and Engineering,Henan University of Technology,Zhengzhou 450001,China; 2.Institute of Remote Sensing and Digital Earth,Chinese Academy of Sciences,Beijing 100094,China)

Greenland; ice-sheet freeze-thaw detection; microwave radiometer; physical model

DOI: 10.13800/j.cnki.xakjdxxb.2017.0622

备注

为提高格陵兰岛冰盖冻融探测精度,在微波辐射计简单冰盖冻融物理模型和SVM(支持向量机)分类方法的基础上,提出了物理模型结合SVM的格陵兰岛冰盖冻融探测新方法,即对微波辐射计简单物理模型得到的冰盖冻融探测结果进行SVM分类,得到格陵兰岛冰盖冻融分布信息。为验证所提方法的可行性与合理性,基于格陵兰岛2013年8月1日的冻融结果和2013年的融化开始时间、持续时间与传统简单物理模型结果进行对比,结果表明:文中所提出的方法得到的格陵兰岛冰盖冻融探测结果更接近于实际的冰盖冻融分布。物理模型结合SVM的冰盖冻融探测方法是有效的、可行的且能有较高的探测精度。

In order to improve snowmelt detection accuracy of Greenland ice sheet,a new snowmelt method is proposed for Greenland ice sheet,which combines physical model with SVM(Support Vector Machine)based on simple snowmelt physical model and SVM classification method for microwave radiometer,that is,the data gotten by the simple snowmelt physical model of the microwave radiometer is classified by SVM model to obtain the information of the snowmelt distribution in Greenland.In order to verify the feasibility and rationality of the proposed method,the results were compared with the traditional simple physical model in Greenland on August 1,2013 and in 2013.The results show that the proposed method is closer to the actual snowmelt distribution.And the proposed method is feasible,effective with higher precision.