基于Contourlet变换域统计模型的SAR图像去噪
|
摘要
合成孔径雷达(Synthetic Aperture Radar,SAR)图像中斑点噪声的存在不利于图像中场景的自动分析和SAR图像的理解,因此斑点噪声的去除对SAR图像的后续处理例如边缘检测、图像分割等是非常重要的。
传统的空域方法和变换域方法对SAR图像进行去噪处理时,会出现细节信息丢失,同质区域内噪声去除不彻底等问题,难以有效地抑制SAR图像的斑点噪声,而多尺度几何分析域统计模型的出现,为SAR图像的去噪处理提供了有效的方法。
本文主要研究了多尺度几何分析工具——Contourlet变换域的统计模型,并将其应用于SAR图像的去噪中,论文主要进行了以下三方面的研究工作:
(1)基于Contourlet域隐马尔可夫树HMT模型和背景隐马尔可夫模型CHMM的SAR图像去噪,该方法以Contourlet系数在邻域内和尺度间都具有很大的依赖性为基础,把HMT和CHMM结合起来建立Contourlet域改进的统计模型对SAR图像进行去噪处理。另外采用Cycle Spinning来抑制伪吉布斯现象,同时利用各向异性扩散来补充细节信息。
(2)基于Contourlet域块隐马尔可夫模型的SAR图像去噪,该方法利用BHMM捕获SAR图像Contourlet系数邻域内的依赖性,进而对SAR图像进行降斑处理。
(3)基于非下采样Contourlet变换NSCT域边缘检测和先验空间约束的SAR图像去噪,该方法先用边缘检测方法对NSCT系数进行分类,然后用BHMM模型对SAR图像进行去噪,最后再利用先验空间约束的方法来对差值图像进行处理,两去噪图像相加得到了较好的去噪结果。
通过对真实SAR图像去噪的实验结果表明,本文所提出的方法取得了比经典空域滤波及其它变换域去噪方法较好的去噪性能。
The presence of speckle noise in SAR images is undesirable, and it makes scene analysis and image understanding very difficult. Thus, speckle reduction is very important for many SAR images processing tasks, for example edge detection and image segmentation.
The traditional denoising methods including airspace domain and transform domain methods don’t work effectively when using for SAR images denoising. When using these methods, some details are lost and the noises of the homogeneous regions can not be effectively disposed of. However, the appearance of statistical models in multiscale geometric transform domain provides kinds of newly developed models for SAR images denoising.
This paper mainly studies Multiscale Geometric Analysis (MGA) tools, i.e. Contourlet and Contourlet based statistical models, on the application of SAR images denoising. Three new methods are proposed in this paper. The main innovative points are as follows:
(1) Contourlet-based Hidden Markov Tree (HMT) Model and Contextual Hidden Markov Model (CHMM) for SAR Images Denoising, this method bases on the fact that Contourlet coefficients bear strong correlations between neighbors and interscales. The CHMM is combined with the HMT to establish statistical models in Contourlet domain. Cycle Spinning will be employed to avoid the artifacts which are caused by translation invariance of the Contourlet transform. Simultaneous, the anisotropic diffusion is used to reinforce the detail information.
(2) Contourlet-based Block Hidden Markov Model (BHMM) for SAR Images Denoising, the BHMM is constructed to characterize the neighbors’dependence of Contourlet coefficients, which is used for SAR images despeckling.
(3) Nonsubsampled Contourlet Transform (NSCT) based Edge Detection and Prior Spatial Constraints for SAR Images Denoising, first using edge detector to label the NSCT coefficients as three classes, then using BHMM for SAR image denoising, at last using prior spatial constraints to deal with difference image, then addition is performed for the two denoising images to get the final denoising result.
The experimental results using real SAR images show that the proposed method outperforms the classical spatial filters and despeckling methods based on others transformed domain.
传统的空域方法和变换域方法对SAR图像进行去噪处理时,会出现细节信息丢失,同质区域内噪声去除不彻底等问题,难以有效地抑制SAR图像的斑点噪声,而多尺度几何分析域统计模型的出现,为SAR图像的去噪处理提供了有效的方法。
本文主要研究了多尺度几何分析工具——Contourlet变换域的统计模型,并将其应用于SAR图像的去噪中,论文主要进行了以下三方面的研究工作:
(1)基于Contourlet域隐马尔可夫树HMT模型和背景隐马尔可夫模型CHMM的SAR图像去噪,该方法以Contourlet系数在邻域内和尺度间都具有很大的依赖性为基础,把HMT和CHMM结合起来建立Contourlet域改进的统计模型对SAR图像进行去噪处理。另外采用Cycle Spinning来抑制伪吉布斯现象,同时利用各向异性扩散来补充细节信息。
(2)基于Contourlet域块隐马尔可夫模型的SAR图像去噪,该方法利用BHMM捕获SAR图像Contourlet系数邻域内的依赖性,进而对SAR图像进行降斑处理。
(3)基于非下采样Contourlet变换NSCT域边缘检测和先验空间约束的SAR图像去噪,该方法先用边缘检测方法对NSCT系数进行分类,然后用BHMM模型对SAR图像进行去噪,最后再利用先验空间约束的方法来对差值图像进行处理,两去噪图像相加得到了较好的去噪结果。
通过对真实SAR图像去噪的实验结果表明,本文所提出的方法取得了比经典空域滤波及其它变换域去噪方法较好的去噪性能。
The presence of speckle noise in SAR images is undesirable, and it makes scene analysis and image understanding very difficult. Thus, speckle reduction is very important for many SAR images processing tasks, for example edge detection and image segmentation.
The traditional denoising methods including airspace domain and transform domain methods don’t work effectively when using for SAR images denoising. When using these methods, some details are lost and the noises of the homogeneous regions can not be effectively disposed of. However, the appearance of statistical models in multiscale geometric transform domain provides kinds of newly developed models for SAR images denoising.
This paper mainly studies Multiscale Geometric Analysis (MGA) tools, i.e. Contourlet and Contourlet based statistical models, on the application of SAR images denoising. Three new methods are proposed in this paper. The main innovative points are as follows:
(1) Contourlet-based Hidden Markov Tree (HMT) Model and Contextual Hidden Markov Model (CHMM) for SAR Images Denoising, this method bases on the fact that Contourlet coefficients bear strong correlations between neighbors and interscales. The CHMM is combined with the HMT to establish statistical models in Contourlet domain. Cycle Spinning will be employed to avoid the artifacts which are caused by translation invariance of the Contourlet transform. Simultaneous, the anisotropic diffusion is used to reinforce the detail information.
(2) Contourlet-based Block Hidden Markov Model (BHMM) for SAR Images Denoising, the BHMM is constructed to characterize the neighbors’dependence of Contourlet coefficients, which is used for SAR images despeckling.
(3) Nonsubsampled Contourlet Transform (NSCT) based Edge Detection and Prior Spatial Constraints for SAR Images Denoising, first using edge detector to label the NSCT coefficients as three classes, then using BHMM for SAR image denoising, at last using prior spatial constraints to deal with difference image, then addition is performed for the two denoising images to get the final denoising result.
The experimental results using real SAR images show that the proposed method outperforms the classical spatial filters and despeckling methods based on others transformed domain.
引文
[1]张澄波.综合孔径雷达原理系统分析与应用.北京：科学出版社，1989.
[2] Henri Maitre，孙洪译.合成孔径雷达图像处理.北京：电子工业出版社，2005.
[3]张直中.机载和星载合成孔径雷达导论.北京：电子工业出版社，2004.
[4]保铮，刑孟道，王彤.雷达成像技术.北京：电子工业出版社，2005.
[5]焦李成，公茂果，王爽等.自然计算、机器学习与图像理解前沿.西安：西安电子科技大学出版社，2008.
[6] J. B. Boisvert, T. J. Pultz, R. J. Brown, et al. Potential of Synthetic Aperture Radarfor Large-Scale Soil Moisture Monitoring: A Review. Canadian Journal ofRemote Sensing. 1995, 22(1):2-13.
[7] Y. Grevier, T.J. Pultz, T. I. Lukowski, et al. Temporal Analysis of ERS-1 SARBackscatter for Hydrology Applications. Canadian Journal of Remote Sensing.1995,22(l):65-76.
[8] Brenner A. R. and Roessing L. Radar Imaging of Urban Areas by Means of VeryHigh-Resolution SAR and Interferometric SAR. IEEE Transactions onGeoscience and Remote Sensing. 2008, 46(10):2971-2982.
[9] Yijun He, Perrie W., Tao Xie, et al. Ocean wave spectra from a linear PolarimetricSAR. IEEE Transactions on Geoscience and Remote Sensing. 2004,42(11):2623-2631.
[10] Moser G. and Karl W. C. Unsupervised Change Detection From MultichannelSAR Data by Markovian Data Fusion. IEEE Transactions on Geoscience andRemote Sensing. 2009, 47(7):2114-2128.
[11] Margarit G., Mallorqui J. J., Rius J.M, et al. On the Usage of GRECOSAR, anOrbital Polarimetric SAR Simulator of Complex Targets, to Vessel ClassificationStudies. IEEE Transactions on Geoscience and Remote Sensing. 2006,44(12):3517-3526.
[12] Gamba P., Dell Acqua F. and Lisini G Change Detection of Multitemporal SARData in Urban Areas Combining Feature-Based and Pixel-Based Techniques.IEEE Transactions on Geoscience and Remote Sensing. 2006, 44(10):2820-2827.
[13] Lance M. Kaplan. Analysis of Multiplicative Speckle Models for Template-BasedSARATR. 2001, 37(4): 1424-1432.
[14] M. Tur, K. C. Chin, and J. W. Goodman. When is speckle noise multiplicative.Applied Optics. 1982, 21(7):1157-1159.
[15] Lee J S. Digital image enhancement and noise filtering by use of local statistics. IEEE Trans. Pattern Analysis and Machine Intelligence. 1980, 2:165-168.
[16] Lopes A., Touzi R., and Nezry E.. Adaptive speckle filters and scence heterogeneity. IEEE Transactions on Geoscience and Remote Sensing. 1990, 28(6):992-1000.
[17] Kuan D. T., Sawchuk A. A., and Strand T. C.et al. Adaptive noise smoothing filter for images with signal-dependent noise. IEEE Trans. Pattern Anal. Mach. Intell. 1985, 7(2):165-177.
[18] Baraldi A, Parmigiani F. A Refined Gamma MAP SAR Speckle Filter with Improved Geometrical Adaptivity. IEEE Trans.Geoscience and Remote Sensing. 1995, 33(5):1245-1257.
[19] Jianwei Ma and Plonka G. Combined Curvelet Shrinkage and Nonlinear Anisotropic Diffusion. IEEE Trans. Image Processing. 2007, 16(9):2198-2206.
[20] Yu Y. J., Acton S. T.. Speckle reduction anisotropic diffusion. IEEE Trans. Image Processing. 2002, 11(11):1260-1270.
[21] Gleich D, Datcu M.Pierce. Wavelet-Based Despeckling of SAR Image Using Gauss-Markov Random Fields. IEEE Trans. Geoscience and Remote Sensing. 2007, 45(12):4127-4143.
[22] Ni W., Guo B. L., and Liu Y.. Speckle reduction algorithm for SAR images using Contourlet transform. J. Inform. and Comput. Sci.2006, 3(1):83-94.
[23] Donoho D L. Denoising by soft-thresholding[J]. IEEE Trans. Information Theory. 1995, 41(3):613-627.
[24] Peng Yuhua. An improved thresholding method in wavelet transform domain for denoising [J].Journal of China Institute of Communications. 2004, 25(8):119-123.
[25] Stéphane Mallat.信号处理的小波导论.杨力华等译.北京:机械工业出版社,2003.
[26] D. L. Donoho, I. M. Johnstone. Adapting to unknown smoothness via wavelet shrinkage. Journal of the American Statistical Association. 1995, 90(432):1200-1224.
[27] Stain C. Estimation of the mean of a multivarials normal distribution. Ann. Stat. 1981, 9:1135-1151.
[28] Chipman H A, Kolaczyk E D, and McCulloch R E. Adaptive Bayesian wavelet shrinkage. J.Ameri.Stat.Assoc. 1997, 92(440):1413-1421.
[29]曹兰英,夏良正.基于贝叶斯估计的非平稳SAR图像自适应小波滤波.航宇学报.2004, 25(3):243-354.
[30] Min Dai, Andrew K. Chan, and Dmitri Logunov. Bayesian Wavelet shrinkage withedge detection for SAR image despeckling. IEEE Trans on Geoscience andRemote Sensing. 2004, 42(8): 1642-1648.
[31] F. Argenti and L. Alparone. Speckle filtering of SAR images in the undecimatedWavelet domain. IEEE Trans on Geoscience and Remote Sensing. 2002,40(ll):2363-2374.
[32] Do M. N and M.. The Contourlet transform: an efficient directionalmultiresolution image representation. IEEE Trans. Image Process. 2005,14(12):2019-2106.
[33] Candes E. J.. Ridgelets: theory and applications. [Ph.D. dissertation], StanfordUniversity, California, USA, 1998.
[34] Candes E. J. and Donoho D. L.. Curvelets-a surprisingly effective nonadaptiverepresentation for objects with edges. L. L. Scgymaker Ed. Curves and Surfaces.Nashville, TN: Vanderbilt University Press. 1999, 1-10.
[35] Meryer F. G and Coifman R. R..Brushlets: a tool for directional image analysisand image compression. Applied and Compuational. Harmonic Analysis. 1997,4(2): 147-187.
[36] Donoho D. L.. Wedgelets: nearly-minimax estimation of edges. Annuals ofStatistics. 1999, 27(3):859-897.
[37] Pennec E. L. and Mallat S.. Sparse geometric image representation with bandelets.IEEE Trans. Image Process. 2005, 14(4):423-438.
[38] Choi H, Baraniuk R G. Multiscale Image Segmentation Using Wavelet-DomainHidden Markov Models. IEEE Trans. Image Processing. 2001, 10(9): 1309-1321.
[39]梅雪，李久贤，夏良正.改进的Contourlet变换域隐马尔可夫树模型.电路与系统学报.2008, 13(4):65-69.
[40] G. Fan, X.-G Xia, Image denoising using a local contextual hidden Markov modelin the wavelet domain. IEEE Signal Processing Letters. 2001, 8(5): 125-128.
[41] Frost V. S., Shanmugan K. S., Holtzman J. C. Edege detection for syntheticaperture radar and other noisy images. Proceedings of. IGARSS'82, Munich,Germany. 1982, FA-2, 4.1-4.9.
[42] Touzi R, Lopes A, and Bousquet P. A statisital and geometrical edge detetor forSAR images. IEEE Transactions on Geoscience and Remote Sensing.1988(26):764-773.
[43]廖增为，宋建社，张宪伟，等.一种基于小波变换的SAR图像多尺度边缘检测融合算法.2005, 22(7):161-163.
[44] Lee J S Jurkevich I. Segmtntation of SAR Images. IEEE Transactions onGeoscience and Remote Sensing. 1989, 27(6):674-680.
[45] Juha A. Karvonen. Batic Sea Ice SAR Segmentation and Classitication UsingModified Pulse-Coupled Neural Networks. IEEE Trans. GARS. 2004,42(7): 1566-1574.
[46] G. Davidson, K. Ouchi. Segmentation of SAR images using multitemporalinformation. Radar, Sonar and Navigation, IEEE Proceedings. 2003,15(10):367-374.
[47]瞿继双，王超，王正志.基于数据融合的遥感图像处理技术.中国图像图像学报.2002, 7(10):985-993.
[48] Cloude S R, E. Pottier. A review of target decomposition theorema in radarpolarimetry. IEEE Transactions on Geoscience and Remote Sensing. 1996,34(2):498-518.
[49] E. J Candes, D. L. Donoho. Curvelets. Tech. Report, Department of Statistics,Stanford University, 1998.
[50] David L. Donoho and Mark R. Duncan. Digital Curvelet Transform: Strategy,Implementation and Experiments. SPIE proceedings series. 1999, 12-29.
[51] E . J. Candes, L. Demanet, D. L. Donoho, et al. Fast Discrete Curvelet Transforms.Applied and Computational Mathematics, Caltech, Pasadena, 2005. CA 91125:1-43.
[52] M. J. Shensa. The discrete Wavelet transform: wedding the atrous and mallatalgorithms. IEEE Trans.on Signal Processing. 1992, 40(10):2464-2482.
[53] P. J. Burt, E. H. Adelson. The Laplacian pyramid as a compact image code. IEEETrans. Commun.. 1983, 31(4):532-540.
[54]焦李成，侯彪，王爽等.图像多尺度几何分析理论与应用.西安：西安电子科技大学出版社，2008.
[55] R. H. Bamberger, M. J. T. Smith. A filter bank for directional decomposition of images: Theory and Design. IEEE Trans. Signal Proc. 1992, 40(4):882-893.
[56] M. N. Do. Directional multiresolution image representations. PhD thesis, Swiss Federal Institute of Technology, Lausanne, Switzerland, December 2001.
[57] Shaowei Dai, Yankui Sun, Xiaolin Tian, et al. Image Denoising Based On Complex Contourlet Transform. Proceedings of 2007 International Conference on Wavelet Analysis and Pattern Recognition. 2007, 1742-1747.
[58] M. Vetterli, C. Herley, Wavelets and filter banks: Theory and design. IEEE Trans.Image Process. 1992, 40(9):2207-2232.
[59] Rabiner L.. A tutorial on hidden Markov models and selected applications in speech recognition. Proceedings of the IEEE. 1989, 77(2):257-286.
[60]孙强.基于统计模型的SAR图像处理和解译.博士学位论文.西安电子科技大学，2007.
[61]李宏东，姚天翔等译.模式分类.模式分类.北京：机械工业出版社，2003.
[62] Forney C. D.. The Viterbi algorithm. Proc. IEEE.1973, 61(3):268-278.
[63] Baum L. E., Petrie T., Soules G., and Weiss N.. Amaximization techniqueoccurring in the statistical analysis of probabilistic functions of Markov chains.Ann. Math. Statist.. 1970, 41(1): 164-171.
[64] Bimes J. A.. A gentle tutorial of the EM algorithm and its application to parameterestimation for Gaussian mixture and hidden Markov models, 1998.
[65] M S Crouse, R D Nowak, R G Baraniuk. Wavelet-based statistical signalprocessing using hidden Markov models. IEEE Trans. Signal Processing. 1998,46(4):886-902.
[66] Crouse M. S., Baraniuk R. G...Contextual hidden Markov models forwavelet-domain signal processing. Proceedings of the 31st Asilomar Conference onSignals, Systems and Computers, Pacific Grove, CA, USA. 1997, 95-100.
[67]焦李成，孙强.多尺度变换域图像的感知与识别：进展和展望.计算机学报.2006, 29(2):177-193.
[68] D.D.-Y. Po, M.N. Do. Directional multiscale modeling of images using the Contourlet transform. IEEE Trans. Image Processing. 2006, 15(6): 1610-1620.
[69] A. Pizurica, W Philips, I. Lemahieu and M. Acheroy. Image de-noising in the wavelet domain using prior spatial constraints. Proc. of the IEEE Int. Conf. on Image Proc. and it's App.. 1999, 7:216-219.
[70] Choi H, Baraniuk R G Multiscale Image Segmentation Using Wavelet-Domain Hidden Markov Models. IEEE Trans. Image Processing. 2001, 10(9): 1309-1321.
[71] Zhiling Long and Nicolas H. Younan. Contourlet Image Modeling with Contextual Hidden Markov Models. Proceedings of the 2006 IEEE Southwest Symposium on Image Analysis and Interpretation. 2006:173-177.
[72]沙宇恒，丛琳，孙强等.基于Contourlet域HMT模型的多尺度图像分割.红外与毫米波学报.2005, 24(6):472-476.
[73] Kang Li, Wei Wang. Jinghuai Gao. Image Denoising Based on Contourlet-Domain HMT Models Using Cycle Spinning. 2008 International Conference on Computer Science and Software Engineering. 2008, 6(12):162-165.
[74] Shaoxiang Hu, Zhiwu Liao. Signal denoising using wavelet packet hiddenMarkov model. Communications, Circuits and Systems, 2004. ICCCAS 2004.2004 International Conference on. 2004, 2(2):751-755.
[75] Zhiling Long, Nicolas H. Younan. Statistical image modeling in the Contourletdomain using contextual hidden Markov models. Elsevier Signal Processing,2009,89:946-951.
[76] T.M .Cover and J.A. Thomas, Elements of Information Theory. New York: Wiley,1991.
[77] R. Moddemeijer. On estimation of entropy and mutual information of continuousdistributional, Signal Process.1989, 16(3):233-246.
[78] J. Beirlant, E.J.Dudewicz, L.Gyofi. Nonparametric entropy estimation: anoverview, IntJ.Math.Statist.Sci. 1997, 17-39.
[79] A. L. Cunha, J. Zhou, and M. N. Do. The nonsubsampled Contourlet transform:Theory, design, and applications. IEEE Trans.on Image Processing. 2006,15(10):3089-3101.
[80] Ridha Touzi, Armand Lopes, and Pierre Bousquet. A statistical and geometricaledge detector for SAR images. IEEE Trans on Geoscience and Remote Sensing.1998,26(6):764-773.
[81] Jose-Gerardo Rosiles, Mark J. T. Smith, and Russell M. Mersereau. SAR imagespeckle removal using Bamberger pyramids. International Conference on ImageProcessing. 2006, 2345-2348.
[82] J. H. McClellan. The design of two-dimensional digital filters by transformation.Proceedings of 7th Annual Princeton Conference on Information Sciences andSystems. 1973,247-251.
[83] W. Sweldens. The lifting scheme: a custom-design construction of biorthogonalWavelets. Applied Computational Harmonic Analysis. 1996, 3(2): 186-200.
[84] S. Mitra, R. Sherwood. Digital ladder networks. IEEE Trans, on AudioElectroacoust. 1973, 21(l):30-36.
[2] Henri Maitre，孙洪译.合成孔径雷达图像处理.北京：电子工业出版社，2005.
[3]张直中.机载和星载合成孔径雷达导论.北京：电子工业出版社，2004.
[4]保铮，刑孟道，王彤.雷达成像技术.北京：电子工业出版社，2005.
[5]焦李成，公茂果，王爽等.自然计算、机器学习与图像理解前沿.西安：西安电子科技大学出版社，2008.
[6] J. B. Boisvert, T. J. Pultz, R. J. Brown, et al. Potential of Synthetic Aperture Radarfor Large-Scale Soil Moisture Monitoring: A Review. Canadian Journal ofRemote Sensing. 1995, 22(1):2-13.
[7] Y. Grevier, T.J. Pultz, T. I. Lukowski, et al. Temporal Analysis of ERS-1 SARBackscatter for Hydrology Applications. Canadian Journal of Remote Sensing.1995,22(l):65-76.
[8] Brenner A. R. and Roessing L. Radar Imaging of Urban Areas by Means of VeryHigh-Resolution SAR and Interferometric SAR. IEEE Transactions onGeoscience and Remote Sensing. 2008, 46(10):2971-2982.
[9] Yijun He, Perrie W., Tao Xie, et al. Ocean wave spectra from a linear PolarimetricSAR. IEEE Transactions on Geoscience and Remote Sensing. 2004,42(11):2623-2631.
[10] Moser G. and Karl W. C. Unsupervised Change Detection From MultichannelSAR Data by Markovian Data Fusion. IEEE Transactions on Geoscience andRemote Sensing. 2009, 47(7):2114-2128.
[11] Margarit G., Mallorqui J. J., Rius J.M, et al. On the Usage of GRECOSAR, anOrbital Polarimetric SAR Simulator of Complex Targets, to Vessel ClassificationStudies. IEEE Transactions on Geoscience and Remote Sensing. 2006,44(12):3517-3526.
[12] Gamba P., Dell Acqua F. and Lisini G Change Detection of Multitemporal SARData in Urban Areas Combining Feature-Based and Pixel-Based Techniques.IEEE Transactions on Geoscience and Remote Sensing. 2006, 44(10):2820-2827.
[13] Lance M. Kaplan. Analysis of Multiplicative Speckle Models for Template-BasedSARATR. 2001, 37(4): 1424-1432.
[14] M. Tur, K. C. Chin, and J. W. Goodman. When is speckle noise multiplicative.Applied Optics. 1982, 21(7):1157-1159.
[15] Lee J S. Digital image enhancement and noise filtering by use of local statistics. IEEE Trans. Pattern Analysis and Machine Intelligence. 1980, 2:165-168.
[16] Lopes A., Touzi R., and Nezry E.. Adaptive speckle filters and scence heterogeneity. IEEE Transactions on Geoscience and Remote Sensing. 1990, 28(6):992-1000.
[17] Kuan D. T., Sawchuk A. A., and Strand T. C.et al. Adaptive noise smoothing filter for images with signal-dependent noise. IEEE Trans. Pattern Anal. Mach. Intell. 1985, 7(2):165-177.
[18] Baraldi A, Parmigiani F. A Refined Gamma MAP SAR Speckle Filter with Improved Geometrical Adaptivity. IEEE Trans.Geoscience and Remote Sensing. 1995, 33(5):1245-1257.
[19] Jianwei Ma and Plonka G. Combined Curvelet Shrinkage and Nonlinear Anisotropic Diffusion. IEEE Trans. Image Processing. 2007, 16(9):2198-2206.
[20] Yu Y. J., Acton S. T.. Speckle reduction anisotropic diffusion. IEEE Trans. Image Processing. 2002, 11(11):1260-1270.
[21] Gleich D, Datcu M.Pierce. Wavelet-Based Despeckling of SAR Image Using Gauss-Markov Random Fields. IEEE Trans. Geoscience and Remote Sensing. 2007, 45(12):4127-4143.
[22] Ni W., Guo B. L., and Liu Y.. Speckle reduction algorithm for SAR images using Contourlet transform. J. Inform. and Comput. Sci.2006, 3(1):83-94.
[23] Donoho D L. Denoising by soft-thresholding[J]. IEEE Trans. Information Theory. 1995, 41(3):613-627.
[24] Peng Yuhua. An improved thresholding method in wavelet transform domain for denoising [J].Journal of China Institute of Communications. 2004, 25(8):119-123.
[25] Stéphane Mallat.信号处理的小波导论.杨力华等译.北京:机械工业出版社,2003.
[26] D. L. Donoho, I. M. Johnstone. Adapting to unknown smoothness via wavelet shrinkage. Journal of the American Statistical Association. 1995, 90(432):1200-1224.
[27] Stain C. Estimation of the mean of a multivarials normal distribution. Ann. Stat. 1981, 9:1135-1151.
[28] Chipman H A, Kolaczyk E D, and McCulloch R E. Adaptive Bayesian wavelet shrinkage. J.Ameri.Stat.Assoc. 1997, 92(440):1413-1421.
[29]曹兰英,夏良正.基于贝叶斯估计的非平稳SAR图像自适应小波滤波.航宇学报.2004, 25(3):243-354.
[30] Min Dai, Andrew K. Chan, and Dmitri Logunov. Bayesian Wavelet shrinkage withedge detection for SAR image despeckling. IEEE Trans on Geoscience andRemote Sensing. 2004, 42(8): 1642-1648.
[31] F. Argenti and L. Alparone. Speckle filtering of SAR images in the undecimatedWavelet domain. IEEE Trans on Geoscience and Remote Sensing. 2002,40(ll):2363-2374.
[32] Do M. N and M.. The Contourlet transform: an efficient directionalmultiresolution image representation. IEEE Trans. Image Process. 2005,14(12):2019-2106.
[33] Candes E. J.. Ridgelets: theory and applications. [Ph.D. dissertation], StanfordUniversity, California, USA, 1998.
[34] Candes E. J. and Donoho D. L.. Curvelets-a surprisingly effective nonadaptiverepresentation for objects with edges. L. L. Scgymaker Ed. Curves and Surfaces.Nashville, TN: Vanderbilt University Press. 1999, 1-10.
[35] Meryer F. G and Coifman R. R..Brushlets: a tool for directional image analysisand image compression. Applied and Compuational. Harmonic Analysis. 1997,4(2): 147-187.
[36] Donoho D. L.. Wedgelets: nearly-minimax estimation of edges. Annuals ofStatistics. 1999, 27(3):859-897.
[37] Pennec E. L. and Mallat S.. Sparse geometric image representation with bandelets.IEEE Trans. Image Process. 2005, 14(4):423-438.
[38] Choi H, Baraniuk R G. Multiscale Image Segmentation Using Wavelet-DomainHidden Markov Models. IEEE Trans. Image Processing. 2001, 10(9): 1309-1321.
[39]梅雪，李久贤，夏良正.改进的Contourlet变换域隐马尔可夫树模型.电路与系统学报.2008, 13(4):65-69.
[40] G. Fan, X.-G Xia, Image denoising using a local contextual hidden Markov modelin the wavelet domain. IEEE Signal Processing Letters. 2001, 8(5): 125-128.
[41] Frost V. S., Shanmugan K. S., Holtzman J. C. Edege detection for syntheticaperture radar and other noisy images. Proceedings of. IGARSS'82, Munich,Germany. 1982, FA-2, 4.1-4.9.
[42] Touzi R, Lopes A, and Bousquet P. A statisital and geometrical edge detetor forSAR images. IEEE Transactions on Geoscience and Remote Sensing.1988(26):764-773.
[43]廖增为，宋建社，张宪伟，等.一种基于小波变换的SAR图像多尺度边缘检测融合算法.2005, 22(7):161-163.
[44] Lee J S Jurkevich I. Segmtntation of SAR Images. IEEE Transactions onGeoscience and Remote Sensing. 1989, 27(6):674-680.
[45] Juha A. Karvonen. Batic Sea Ice SAR Segmentation and Classitication UsingModified Pulse-Coupled Neural Networks. IEEE Trans. GARS. 2004,42(7): 1566-1574.
[46] G. Davidson, K. Ouchi. Segmentation of SAR images using multitemporalinformation. Radar, Sonar and Navigation, IEEE Proceedings. 2003,15(10):367-374.
[47]瞿继双，王超，王正志.基于数据融合的遥感图像处理技术.中国图像图像学报.2002, 7(10):985-993.
[48] Cloude S R, E. Pottier. A review of target decomposition theorema in radarpolarimetry. IEEE Transactions on Geoscience and Remote Sensing. 1996,34(2):498-518.
[49] E. J Candes, D. L. Donoho. Curvelets. Tech. Report, Department of Statistics,Stanford University, 1998.
[50] David L. Donoho and Mark R. Duncan. Digital Curvelet Transform: Strategy,Implementation and Experiments. SPIE proceedings series. 1999, 12-29.
[51] E . J. Candes, L. Demanet, D. L. Donoho, et al. Fast Discrete Curvelet Transforms.Applied and Computational Mathematics, Caltech, Pasadena, 2005. CA 91125:1-43.
[52] M. J. Shensa. The discrete Wavelet transform: wedding the atrous and mallatalgorithms. IEEE Trans.on Signal Processing. 1992, 40(10):2464-2482.
[53] P. J. Burt, E. H. Adelson. The Laplacian pyramid as a compact image code. IEEETrans. Commun.. 1983, 31(4):532-540.
[54]焦李成，侯彪，王爽等.图像多尺度几何分析理论与应用.西安：西安电子科技大学出版社，2008.
[55] R. H. Bamberger, M. J. T. Smith. A filter bank for directional decomposition of images: Theory and Design. IEEE Trans. Signal Proc. 1992, 40(4):882-893.
[56] M. N. Do. Directional multiresolution image representations. PhD thesis, Swiss Federal Institute of Technology, Lausanne, Switzerland, December 2001.
[57] Shaowei Dai, Yankui Sun, Xiaolin Tian, et al. Image Denoising Based On Complex Contourlet Transform. Proceedings of 2007 International Conference on Wavelet Analysis and Pattern Recognition. 2007, 1742-1747.
[58] M. Vetterli, C. Herley, Wavelets and filter banks: Theory and design. IEEE Trans.Image Process. 1992, 40(9):2207-2232.
[59] Rabiner L.. A tutorial on hidden Markov models and selected applications in speech recognition. Proceedings of the IEEE. 1989, 77(2):257-286.
[60]孙强.基于统计模型的SAR图像处理和解译.博士学位论文.西安电子科技大学，2007.
[61]李宏东，姚天翔等译.模式分类.模式分类.北京：机械工业出版社，2003.
[62] Forney C. D.. The Viterbi algorithm. Proc. IEEE.1973, 61(3):268-278.
[63] Baum L. E., Petrie T., Soules G., and Weiss N.. Amaximization techniqueoccurring in the statistical analysis of probabilistic functions of Markov chains.Ann. Math. Statist.. 1970, 41(1): 164-171.
[64] Bimes J. A.. A gentle tutorial of the EM algorithm and its application to parameterestimation for Gaussian mixture and hidden Markov models, 1998.
[65] M S Crouse, R D Nowak, R G Baraniuk. Wavelet-based statistical signalprocessing using hidden Markov models. IEEE Trans. Signal Processing. 1998,46(4):886-902.
[66] Crouse M. S., Baraniuk R. G...Contextual hidden Markov models forwavelet-domain signal processing. Proceedings of the 31st Asilomar Conference onSignals, Systems and Computers, Pacific Grove, CA, USA. 1997, 95-100.
[67]焦李成，孙强.多尺度变换域图像的感知与识别：进展和展望.计算机学报.2006, 29(2):177-193.
[68] D.D.-Y. Po, M.N. Do. Directional multiscale modeling of images using the Contourlet transform. IEEE Trans. Image Processing. 2006, 15(6): 1610-1620.
[69] A. Pizurica, W Philips, I. Lemahieu and M. Acheroy. Image de-noising in the wavelet domain using prior spatial constraints. Proc. of the IEEE Int. Conf. on Image Proc. and it's App.. 1999, 7:216-219.
[70] Choi H, Baraniuk R G Multiscale Image Segmentation Using Wavelet-Domain Hidden Markov Models. IEEE Trans. Image Processing. 2001, 10(9): 1309-1321.
[71] Zhiling Long and Nicolas H. Younan. Contourlet Image Modeling with Contextual Hidden Markov Models. Proceedings of the 2006 IEEE Southwest Symposium on Image Analysis and Interpretation. 2006:173-177.
[72]沙宇恒，丛琳，孙强等.基于Contourlet域HMT模型的多尺度图像分割.红外与毫米波学报.2005, 24(6):472-476.
[73] Kang Li, Wei Wang. Jinghuai Gao. Image Denoising Based on Contourlet-Domain HMT Models Using Cycle Spinning. 2008 International Conference on Computer Science and Software Engineering. 2008, 6(12):162-165.
[74] Shaoxiang Hu, Zhiwu Liao. Signal denoising using wavelet packet hiddenMarkov model. Communications, Circuits and Systems, 2004. ICCCAS 2004.2004 International Conference on. 2004, 2(2):751-755.
[75] Zhiling Long, Nicolas H. Younan. Statistical image modeling in the Contourletdomain using contextual hidden Markov models. Elsevier Signal Processing,2009,89:946-951.
[76] T.M .Cover and J.A. Thomas, Elements of Information Theory. New York: Wiley,1991.
[77] R. Moddemeijer. On estimation of entropy and mutual information of continuousdistributional, Signal Process.1989, 16(3):233-246.
[78] J. Beirlant, E.J.Dudewicz, L.Gyofi. Nonparametric entropy estimation: anoverview, IntJ.Math.Statist.Sci. 1997, 17-39.
[79] A. L. Cunha, J. Zhou, and M. N. Do. The nonsubsampled Contourlet transform:Theory, design, and applications. IEEE Trans.on Image Processing. 2006,15(10):3089-3101.
[80] Ridha Touzi, Armand Lopes, and Pierre Bousquet. A statistical and geometricaledge detector for SAR images. IEEE Trans on Geoscience and Remote Sensing.1998,26(6):764-773.
[81] Jose-Gerardo Rosiles, Mark J. T. Smith, and Russell M. Mersereau. SAR imagespeckle removal using Bamberger pyramids. International Conference on ImageProcessing. 2006, 2345-2348.
[82] J. H. McClellan. The design of two-dimensional digital filters by transformation.Proceedings of 7th Annual Princeton Conference on Information Sciences andSystems. 1973,247-251.
[83] W. Sweldens. The lifting scheme: a custom-design construction of biorthogonalWavelets. Applied Computational Harmonic Analysis. 1996, 3(2): 186-200.
[84] S. Mitra, R. Sherwood. Digital ladder networks. IEEE Trans, on AudioElectroacoust. 1973, 21(l):30-36.