基于MAR-MRF的SAR图像分割方法

刘爱平; 付琨; 尤红建; 刘忠

doi:10.3724/SP.J.1146.2008.01543

基于MAR-MRF的SAR图像分割方法

doi: 10.3724/SP.J.1146.2008.01543

刘爱平^{①②;付琨},
付琨,
尤红建,
刘忠

基金项目:

中国科学院电子学研究所创新项目资助课题

计量
- 文章访问数: 3771
- HTML全文浏览量: 115
- PDF下载量: 976
- 被引次数: 0
出版历程
- 收稿日期: 2008-11-24
- 修回日期: 2009-05-20
- 刊出日期: 2009-11-19

SAR Image Segmentation Based on Multiscale AutoRegressive and Markov Random Field Models

摘要

摘要: 该文提出了一种基于多尺度自回归模型和马尔科夫随机场的SAR图像分割算法。算法引入多尺度自回归模型，建立层与层之间以及相邻层的像素点之间的数学关系，并将此模型与马尔科夫分割算法结合，实现了更为合理的多尺度分割策略。通过相邻尺度的依赖关系及同一尺度空间的马尔可夫性，使用多尺度自回归模型的预测结果来引导精细尺度图像分割，不仅使得最细尺度下的分割迭代次数减少；而且去除了最细尺度下多余的误分类斑块；同时还能够分割出清晰、平滑的目标边界，实现了较满意的SAR图像分割。
- SAR图像处理;多尺度自回归;马尔科夫随机场;多尺度分割;吉布斯随机场
Abstract: A method of SAR image segmentation based on Multiscale AutoRegressive and Markov Random Field (MAR-MRF) models is presented. MAR models is used to establish mathematic relationship among different image layers, and is combined with Markov Random Field (MRF) segment models. This method takes into account the dependence of neighbor layers Markov property of the same layer, and uses forecasting result of the MAR models to direct the fine layer segmentation. Experimental results on SAR image show that this method reduces the iterative times of segmentation and inaccuracy classify blocks, and gets clear and smooth object border.

HTML全文

参考文献(1)

Xue Xiao-rong.[J].Zhang Yan-ning, Zhao Rong-chun, DuanFeng, and Chen Yi. A new method of SAR imagesegmentation based on neural network. ComputerIntelligence and Multimedia Application, ICCIMA 2003,Xian, China.2003,:-[2]Cao Lan-ying, Zhang Kun-hui, and Xia Liang-zheng. SARImage segmentation by 2-Dfussy entropy. Geoscience andRemote Sensing Symposium, 2004. IGARSS apos, 04processing. Anchorage, Alaska, USA, 2004, Vol.6: 3798-3801.[3]Fjortoft R, Lopes A, Marthon P, and Cubero-Castan E. Anoptimal multiedge detector for SAR image segmentation[J].IEEE Transactions on Geoscience and Remote Sensing.1998,36(3):793-802[4]Chen Ju and Moloney C R. An edge-enhanced segmentationmethod for SAR images. IEEE 1997 Canadiam Conference onElectrical and Computer Engineering, St. Johns, Nfld.,Canada, 1997, Vol.2: 599-602.[5]Martinez P, Schertzer D, and Pham K K. Texturemodelisation by multifractal processes for SAR imagesegmentation. IEE conference of Radar 97, Edinburgh, UK.1997: 135-139.[6]Gan Du and Tat Soon Yeo. A novel lacunarity estimationmethod applied to SAR image segmentation[J].IEEETransactions on Geoscience and Remote Sensing.2002,40(12):2687-2691[7]Felzenszwalb P, McAllester D, and Ramanan D. Adiscriminatively trained, multiscale, deformable part model[C]. IEEE conference on Computervision and patternrecognition 2008, CVPR 2008, Anchorage, Alaska, 2008: 1-8.[8]Meirav Galun, Ronen Basri, and Achi Brandt. Multiscaleedge detection and fiber enhancement using differences oforiented means [C]. IEEE 11th International Conference onComputer Vision 2007, ICCV 2007, Rio de Janeiro, Brazil,2007: 1-8.[9]Larlus D and Jurie F. Combining appearance models andMarkov random fields for category level objectsegmentation[C]. IEEE conference on Computervision andpattern recognition 2008, CVPR 2008, Anchorage, Alaska,2008: 1-7.Guo Yan-lin, Cen Rao, Samarasekera S, Kim J, and Kumar R.Matching vehicles under large pose transformations usingapproximate 3D models and piecewise MRF model[C]. IEEEconference on Computervision and pattern recognition 2008,CVPR 2008, Anchorage, Alaska, 2008: 1-8.[10]句彦伟, 田铮, 纪建. SAR图像无监督分割的空间变化混合MAR模型方法[J]. 计算机学报, 2006, 29(2): 331-336.Ju Yan-wei, Tian Zheng, and Ji Jian. SAR imageryunsupervised segmentation based on spatially variantmixture multiscale autoregressive model [J]. Chinese Journalof Computers, 2006, 29(2): 331-336.[11]吴永辉, 计科峰, 李禹, 郁文贤. 基于Wishart分布和MRF的多视全极化SAR 图像分割[J]. 电子学报, 2007, 35(12):2302-2306.Wu Yong-hui, Ji Ke-feng, Li Yu, and Yu Wen-xian.Segmentation of multi-look fully polarimetric SAR imagesbased on Wishart distribution and MRF [J]. Acta ElectronicaSinica, 2007, 35(12): 2302-2306.

施引文献

资源附件(0)

访问统计