基于EM算法的G0分布参数最大似然估计

doi:10.3969/j.issn.0372-2112.2013.01.031

电子学报 ›› 2013, Vol. 41 ›› Issue (1): 178-184.DOI: 10.3969/j.issn.0372-2112.2013.01.031

基于EM算法的G0分布参数最大似然估计

周鑫

南京航空航天大学自动化学院,江苏南京 210016

收稿日期:2012-05-14 修回日期:2012-08-30 出版日期:2013-01-25
- 作者简介:
- 周鑫 男,1980年1月出生,江苏丹阳人.南京航空航天大学自动化学院副教授.2008年在弗吉尼亚大学获博士学位.主要研究方向为目标检测和识别,图像处理和分析. E-mail:xzhou@nuaa.edu.cn
- 基金资助:
- 国家自然科学基金 (No.61102138); 南京航空航天大学基本科研业务费专项基金 (No.201090)

An EM Algorithm Based Maximum Likelihood Parameter Estimation Method for the G0 Distribution

ZHOU Xin

College of Automation Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu 210016, China

Received:2012-05-14 Revised:2012-08-30 Online:2013-01-25 Published:2013-01-25
- Supported by:
- National Natural Science Foundation of China (No.61102138); Fundamental Research Funds for Nanjing University of Aeronautics and Astronautics (No.201090)

摘要/Abstract

摘要： G0分布是目前合成孔径雷达(Synthetic Aperture Radar,SAR)图像数据建模的一个重要模型,建模能力强、实用性好,受到了广泛的关注.G0分布的应用离不开准确有效的参数估计,而由于G0分布表达式复杂,统计意义上最优的最大似然估计法一直没能用在G0分布上.本文首先给出了一种新的方式来推导得出G0分布,在此基础上,采用最大期望(Expectation Maximization,EM)算法为G0分布给出一种有效的最大似然参数估计方法.文中的方法与现有的G0分布参数估计方法通过实验进行了比较,实验结果充分证明了所提方法的有效性.

关键词: SAR图像, G0分布, EM算法, 最大似然估计

Abstract: As an important model for modeling synthetic aperture radar (SAR) image data,G0 distribution has strong modeling ability and good practicability,and therefore draws extensive attentions around the world.The applications of G0 distribution require accurate and effective parameter estimations,and maximum likelihood estimator,which is statistically optimal,has not been applied for the G0 distribution due to the complexity of G0 distribution.In this paper,a new derivation of G0 distribution is first given,based on which,a maximum likelihood parameter estimation method using expectation maximization (EM) algorithm is proposed for the G0 ditribution.The proposed method is compared with other G0 parameter estimation methods through extensive experiments,and the results show the effectiveness of the proposed method.

Key words: SAR image, G0 distribution, EM algorithm, maximum likelihood estimation

中图分类号:

TN957

周鑫. 基于EM算法的G0分布参数最大似然估计[J]. 电子学报, 2013, 41(1): 178-184.

ZHOU Xin. An EM Algorithm Based Maximum Likelihood Parameter Estimation Method for the G0 Distribution[J]. Acta Electronica Sinica, 2013, 41(1): 178-184.

参考文献

[1] Oliver C J,Quegan S.Understanding Synthetic Aperture Radar Images [M].London:Artech House,1998.
[2] Novak L M,Owirka G J,Brower W S,Weaver A L.The automatic target-recognition system in SAIP [J].The Lincoln Laboratory Journal,1997,10(2):187-202.
[3] 万朋,王建国,黄顺吉.SAR图像目标综合检测方法[J].电子学报,2001,29(3):323-325.Wan Peng,Wang Jian-guo,Huang Shun-ji.A synthesis method for SAR image target detection [J].Acta Electronica Sinica,2001,29(3):323-325.(in Chinese)
[4] 付琨,孙真真,吴一戎.基于Beta-Prime统计模型和QGD分类器的SAR图像地物分类方法[J].电子学报,2003,31(12A):2163-2166.Fu Kun,Sun Zhen-zhen,Wu Yi-rong.The SAR image terrain classification algorithm combining quadratic Gamma discrimination classifier and Beta-Prime statistic model [J].Acta Electronica Sinica,2003,31(12A):2163-2166.(in Chinese)
[5] 刘向阳,许稼,彭应宁.极不均匀合成孔径雷达杂波建模及恒虚警检测[J].电子学报,2007,35(9):1617-1621.Liu Xiang-yang,Xu Jia,Peng Ying-ning.Model validation of the extremely heterogeneous SAR clutter and its CFAR detection [J].Acta Electronica Sinica,2007,35(9):1617-1621.(in Chinese)
[6] 高贵.SAR图像统计建模研究综述[J].信号处理,2009,25(8):1270-1278.Gao Gui.Review on the statistical modeling of SAR image [J].Signal Processing,2009,25(8):1270-1278.(in Chinese)
[7] 高贵,鲁敏,黄纪军,匡纲要,李德仁.高分辨率SAR图像中杂波的统计特性分析[J].信号处理,2008,24(4):648-654.Gao Gui,Lu Min,Huang Ji-jun,Kuang Gang-yao,Li De-ren.Statistical analysis of clutter in high resolution SAR images [J].Signal Processing,2008,24(4):648-654.(in Chinese)
[8] Frery A C,Muller H,Freitas C C,et al.A model for extremely heterogeneous clutter [J].IEEE Transactions on GRS,1997,35(3):648-659.
[9] 周晓光.极化SAR图像分类方法研究[D].湖南长沙:国防科学技术大学,2008,10.Zhou Xiao-guang.Polarimetric SAR Image Classification [D].Changsha,Hunan:National University of Defense Technology,2008,10.(in Chinese)
[10] Mejail M E,Jacobo-berlles J C,Frery A C,et al.Classification of SAR images using a general and tractable multiplicative model [J].International Journal of Remote Sensing,2003,24(18):3565-3582.
[11] 周晓光,贺志国,匡纲要,万建伟.基于极化G0分布和MRF的多视PolSAR图像迭代分类方法[J].宇航学报,2009,30(1):276-281.Zhou Xiao-guang,He Zhi-guo,Kuang gang-yao,Wan Jian-wei.Iterative classification for multi-look PolSAR images based on polarimetric G0 distribution and MRF [J].Journal of Astronautics,2009,30(1):276-281.(in Chinese)
[12] Marques R C P,Medeiros F N S,Nobre J.SAR image segmentation based on level set approach and G_A^0 model [J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(10):2046-2057.
[13] Freitas C C,Frery A C,Correia A H.The polarimetric G distribution for SAR data analysis [J].Environmetrics,2005,16(1):13-31.
[14] 时公涛,高贵,周晓光,匡纲要,蒋咏梅.基于Mellin变换的G0分布参数估计方法[J].自然科学进展,2009,19(6):677-690.Shi Gong-tao,Gao gui,Zhou Xiao-guang,Kuang Gang-yao,Jiang Yong-mei.A parameter estimation method for G⁰ distribution based on Mellin transform [J].Progress in Natural Science,2009,19(6):677-690.(in Chinese)
[15] Tison C,Nicolas J-M,Tupin F,Maitre H.A new statistical model for Markovian classification of urban areas in high-resolution SAR images [J].IEEE Transactions on Geoscience and Remote Sensing,2004,42(10):2046-2057.
[16] Berger J O.Statistical Decision Theory and Bayesian Analysis (2nd edition) [M].New York:Springer,1985.
[17] Roberts W J J,Furui S.Maximum likelihood estimation of K-distribution parameters via the expectation-maximization algorithm [J].IEEE Transactions on Signal Processing,2000,48(12):3303-3306.
[18] Dempster A P,Laird N M,Rubin D B.Maximum likelihood from incomplete data via the EM algorithm (with discussion) [J].Journal of the Royal Statistical Society,Series B,1977,39(1):1-38.
[19] Gradshteyn I S,Ryzhik I M.Table of Integrals,Series,and Products (seventh edition) [M].Oxford,UK:Academic Press,2007.
[20] DeVore M D,O'Sullivan J A,Montagnino L J.Conditionally Gamma and K distribution models for SAR ATR [A].Proceedings of Algorithms for Synthetic Aperture Radar Imagery IX [C].Orlando:SPIE,2002.27-37.

基于EM算法的G0分布参数最大似然估计

An EM Algorithm Based Maximum Likelihood Parameter Estimation Method for the G0 Distribution

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	赵琰, 赵凌君, 匡纲要. 基于注意力机制特征融合网络的SAR图像飞机目标快速检测[J]. 电子学报, 2021, 49(9): 1665-1674.
[2]	魏志强, 毕海霞, 刘霞. 基于深度卷积神经网络的图上半监督极化SAR图像分类算法[J]. 电子学报, 2020, 48(1): 66-74.
[3]	王超, 刘淇, 陈恩红, 黄振亚, 朱天宇, 苏喻, 胡国平. 面向大规模认知诊断的DINA模型快速计算方法研究[J]. 电子学报, 2018, 46(5): 1047-1055.
[4]	钟志明, 谭坚文, 叶方伟, 刘阳, 曾德平, 吴明达. 基于统计分析的多协议单片机数据通信系统EFT抗扰度研究[J]. 电子学报, 2018, 46(2): 393-400.
[5]	唐波, 张玉, 张浩. 基于交替投影的MIMO雷达信号盲分离算法研究[J]. 电子学报, 2017, 45(9): 2092-2097.
[6]	王鼎, 张刚. 一种基于窄带信号多普勒频率测量的运动目标直接定位方法[J]. 电子学报, 2017, 45(3): 591-598.
[7]	王峰, 尤红建, 傅兴玉, 许宁. 应用于多源SAR图像匹配的级联SIFT算法[J]. 电子学报, 2016, 44(3): 548-554.
[8]	詹曙, 王俊, 杨福猛, 方琪. 基于Gabor特征和字典学习的高斯混合稀疏表示图像识别[J]. 电子学报, 2015, 43(3): 523-528.
[9]	熊坤来, 刘章孟, 柳征, 姜文利, 汪华兴. 基于EM算法的宽带信号DOA估计及盲分离[J]. 电子学报, 2015, 43(10): 2028-2033.
[10]	杨勇, 张冬玲, 彭华, 涂世龙. 单通道高阶调制线性混合信号的频偏估计[J]. 电子学报, 2015, 43(1): 160-165.
[11]	张永梅, 季艳, 马礼, 张睿, 李洁琼, 熊焰. 遥感图像建筑物识别及变化检测方法[J]. 电子学报, 2014, 42(4): 653-657.
[12]	王先文, 陈锋, 程智, 杜耀华, 暴洪涛, 吴太虎. 基于偏斜t混合模型的流式数据自动聚类方法研究[J]. 电子学报, 2014, 42(12): 2527-2535.
[13]	项德良, 粟毅, 赵凌君, 唐涛, 陆军. 一种基于局部梯度比率特征度量SAR图像相似性的新方法[J]. 电子学报, 2014, 42(1): 9-13.
[14]	刘涛, 崔浩贵, 高俊. Wishart分布矩阵行列式值的统计特性及其在参数估计中的应用[J]. 电子学报, 2013, 41(6): 1231-1237.
[15]	蒋李兵, 王壮, 雷琳, 郭炜炜, 郁文贤. 基于模型的单幅高分辨SAR图像建筑物高度反演方法[J]. 电子学报, 2012, 40(6): 1086-1091.