段盈宏,
谭维贤,,
徐伟
内蒙古工业大学信息工程学院 呼和浩特 010051
内蒙古自治区雷达技术与应用重点实验室 呼和浩特 010051
基金项目:国家自然科学基金(61631011),内蒙古科技重大专项(2019ZD022),内蒙古科技计划项目(2019GG139),内蒙古创新引导项目(KCBJ2017, KCBJ2018014)
详细信息
作者简介:黄平平(1978–),男,山东海阳人,博士,教授。2010年于中国科学院电子学研究所获工学博士学位,现任内蒙古工业大学信息工程学院副院长,自治区雷达技术与应用重点实验室主任,“草原英才”创新团队负责人,全国“工人先锋号”负责人。兼任中央军委装备发展部某专家组成员、中国电子学会信号处理分会常务委员、中国电子教育学会研究生教育分会理事。入选国家“百千万人才工程”、国家有突出贡献中青年专家、自治区“草原英才”、自治区自然科学基金****。研究方向为新体制雷达系统、雷达信号处理和微波遥感应用。E-mail: hpp@imut.edu.cn
段盈宏(1995–),男,河北石家庄人,现于内蒙古工业大学信息工程学院雷达技术与应用重点实验室攻读硕士学位,主要研究方向为多源遥感图像处理方法研究。E-mail: dy_h1995@163.com
谭维贤(1981–),男,博士,教授,硕士生导师。研究方向为微波二维/三维成像技术、微变监测雷达和微波遥感应用等。E-mail: wxtan@imut.edu.cn
徐伟:徐 伟(1983–),男,博士,教授,硕士生导师。2011年获中国科学院电子学研究所工学博士学位;2011—2017年中国科学院电子学研究所副研究员;现为内蒙古工业大学信息工程学院教师。目前主要研究方向为新体制雷达系统、雷达信号处理和微波遥感应用。E-mail:xuwei1983@imut.edu.cn
通讯作者:黄平平 hpp@imut.edu.cn
谭维贤 wxtan@imut.edu.cn
责任主编:李刚 Corresponding Editor: LI Gang中图分类号:TP753
计量
文章访问数:636
HTML全文浏览量:319
PDF下载量:108
被引次数:0
出版历程
收稿日期:2020-08-25
修回日期:2020-10-27
网络出版日期:2020-11-10
Change Detection Method Based on Fusion Difference Map in Flood Disaster
HUANG Pingping,,DUAN Yinghong,
TAN Weixian,,
XU Wei
College of Information Engineering, Inner Mongolia University of Technology, Hohhot 010051, China
Inner Mongolia Key Laboratory of Radar Technology and Application, Hohhot 010051, China
Funds:The National Natural Science Foundation of China (61631011), Major Science and Technology Project of Inner Mongolia (2019ZD022), Planned Project of Science and Technology of Inner Mongolia (2019GG139), Innovation Guidance Project of Inner Mongolia (KCBJ2017, KCBJ2018014)
More Information
Corresponding author:HUANG Pingping, hpp@imut.edu.cn;TAN Weixian, wxtan@imut.edu.cn
摘要
摘要:由于洪灾区域的地物散射特性受环境影响会发生改变,在对该区域合成孔径雷达(SAR)图像进行变化检测时会使检测结果的错误率提高,而且用单一方法得到的差异图变化检测结果精度较低。针对上述问题,该文提出一种基于融合差异图的变化检测方法,该方法通过构造基于改进相对熵与均值比的融合差异图,综合了熵值差异图的区域敏感性和均值差异图的区域保持性的优势。首先,利用皮尔逊相关系数对模糊局部信息C均值聚类(FLICM)方法的初始聚类结果进行二次分类,再将二次分类结果作为图像初始分割,最后利用迭代条件模型和马尔科夫随机场(ICM-MRF)获得图像的最终分割结果。为了验证所提方法的有效性,该文使用瑞士Bern地区在1999年4月和5月的ERS-2遥感数据以及加拿大Ottawa地区在1997年5月和8月的Radarsat遥感数据进行实验,并用该方法对中国鄱阳湖地区2020年6月和7月的Sentinel-1-A遥感数据进行了洪灾检测实验,估计了鄱阳湖附近区域洪灾前后的受灾范围和变化趋势。实验结果表明该文算法总体检测误差较低,一定程度上降低了检测结果的错误率,提高了检测结果的精度。
关键词:SAR图像/
变化检测/
无监督/
改进相对熵/
迭代条件模型和马尔科夫随机场
Abstract:Due to the influence of the environment on the scattering characteristics of ground objects in flooded areas, the false error rate of the detection results increases when performing change detection on Synthetic Aperture Radar (SAR) images of these areas, which reduces the accuracy of the results obtained for the difference map. To solve this problem, in this paper, we propose a change-detection method based on a fusion difference map. This method combines the regional sensitivity of the entropy difference map with the regional retention of the mean difference map to construct a fusion difference map based on an improved relative entropy and mean value ratio. First, the initial clustering results of the fuzzy local information C-means clustering method are classified by their Pearson correlation coefficients, and second, the secondary classification results are used for the initial image segmentation. Third, the final segmentation results are obtained using the iterative condition model and Markov random field. To verify the flood-disaster-detection performance of the proposed method, we used the second of Europe Remote-Sensing (ERS-2) Satellite data obtained for the Bern area in Switzerland in April and May 1999 and Radarsat remote-sensing data for the Ottawa region in Canada in May and August 1997. We also applied the proposed method to data obtained for the Poyang Lake region of China in June and July 2020, and estimated the disaster area and change trend before and after the flood in Poyang Lake. The experimental results show that the algorithm had a low overall detection error, the false error rate of the detection results were somewhat reduced, and the accuracy of the detection results was improved.
Key words:SAR image/
Change detection/
Unsupervised/
Improve relative entropy/
Iterative Condition Model and Markov Random Field (ICM-MRF)
PDF全文下载地址:
https://plugin.sowise.cn/viewpdf/198_2035a9f9-389e-4857-8070-ad66e280a6a9_R20118