An Improved Markov Random Field Classification Approach for Hyperspectral Data Based on Efficient Belief Propagation
CAO Yang1,2 ,ZHAO Hui-Jie2 ,HUANG Si-Niu1 ,LI Na2,ZHANG Pei1
1The RD and Application Center of Command Automation Technology,The 4th Institute of CASIC,Beijing 102308 2Beijing University of Aeronautics and Astronautics,Key Laboratory of Precision Opto-mechatronics Technology of Education Ministry,Beijing 100191
Abstract Aiming at the problems of imprecise class conditional probability (CCP) estimation and heavy computational cost for the global energy minimum in Markov random field (MRF) based classification algorithm,an improved MRF approach based on efficient belief propagation (EBP) is developed for land-cover classification of hyperspectral data. The estimation accuracy of the CCP is improved by the probabilistic support vector machine (PSVM) algorithm using spectral information of pixels,then the spatial correlation information is introduced by the MRF classification algorithm,thus the spectral information and spatial information is combined effectively. Moreover,an EBP optimization algorithm is developed,by which the computational cost is reduced and the classification accuracy is improved. The experimental results show that the proposed approach is effective. The average classification accuracy is up to 95.78%,Kappa coefficient is 93.34%,and the computational time of EBP is about 25% of that by belief propagation algorithm. Therefore,the proposed approach is valuable in land-cover classification application for hyperspectral data with low computational cost and high classification accuracy.
CAO Yang,ZHAO Hui-Jie,HUANG Si-Niu等. An Improved Markov Random Field Classification Approach for Hyperspectral Data Based on Efficient Belief Propagation[J]. , 2014, 27(3): 248-255.
[1] Tong Q X,Zhang B,Zheng L F. Hyperspectral Remote Sensing-Theory,Technique and Application. Beijing,China: Higher Education Press,2006 (in Chinese) (童庆禧,张 兵,郑兰芬.高光谱遥感——原理、技术与应用.北京:高等教育出版社,2006) [2] Dong C. Research on Key Techniques of Supervised Classification Algorithm of Hyperspectral Remote Sensing Image. Ph.D Dissertation. Beijing,China: Beijing University of Aeronautics and Astronautics,2010 (in Chinese) (董 超.高光谱遥感影像监督分类算法关键技术研究.博士学位论文.北京:北京航空航天大学,2010) [3] Liu X. Automatic Target Detection on Hyperspectral Imagery Based on Transformation of Spectral Dimensions. Ph.D Dissertation. Beijing,China: Institute of Remote Sensing Applications,Chinese Academy of Sciences,2008 (in Chinese) (刘 翔.基于光谱维变换的高光谱图像目标探测研究.博士学位论文.北京:中国科学院遥感应用研究所,2008) [4] Theiler J,Scovel C,Wohlberg B,et al. Elliptically Contoured Distributions for Anomalous Change Detection in Hyperspectral Imagery. IEEE Geoscience and Remote Sensing Letters,2010,7(2): 271-275 [5] CristianinI N,Shawe-Taylor J. An Introduction to Support Vector Machines and Other Kernel-Based Learning Methods. Cambridge,UK: Cambridge University Press,2000 [6] Tan K,Du P J. Wavelet Support Vector Machines Based on Reproducing Kernel Hilbert Space for Hyperspectral Remote Sensing Image Classification. Acta Geodaetica et Cartographica Sinica,2011,40(2): 142-147 (in Chinese) (谭 琨,杜培军.基于再生核Hilbert 空间小波核函数支持向量机的高光谱遥感影像分类.测绘学报,2011,40(2): 142-147) [7] Tso B,Mather P M. Classification Methods for Remotely Sensed Data. Boca Raton,USA: CRC Press,2001 [8] Chen J,Deng M,Xiao P F,et al. Object-Oriented Classification of High Resolution Imagery Combining Support Vector Machine with Granular Computing. Acta Geodaetica et Cartographica Sinica,2011,40(2): 135-141 (in Chinese) (陈杰,邓敏,肖鹏峰,等.结合支持向量机与粒度计算的高分辨率遥感影像面向对象分类.测绘学报,2011,40(2): 135-141) [9] Gao H Z,Wan J W,Nian Y J,et al. Hyperspectral Image Classification Algorithm Based on Spectral-Spatial Hybrid Features and SVM. Journal of Astronautics,2011,32(4): 917-921 (in Chinese) (高恒振,万建伟,粘永健,等.一种基于谱域－空域组合特征支持向量机的高光谱图像分类算法.宇航学报,2011,32(4): 917-921) [10] Gao H Z,Wan J W,Wang L B,et al. Research on Classification Technique for Hyperspectral Imagery Based on Spectral-Spatial Composite Kernels. Signal Processing,2011,27(5): 648-652 (in Chinese) (高恒振,万建伟,王力宝,等.基于谱域-空域组合核函数的高光谱图像分类技术研究.信号处理,2011,27(5): 648-652) [11] Fauvel M,Benediktsson J A,Chanussot J,et al. Spectral and Spatial Classification of Hyperspectral Data Using SVMs and Morphological Profiles.IEEE Trans on Geoscience and Remote Sensing,2008,46(11): 3804-3814 [12] Rottensteiner F,Trinder J,Clode S,et al. Using the Dempster-Shafer Method for the Fusion of LIDAR Data and Multi-Spectral Images for Building Detection. Information Fusion,2005,6(4): 283-300 [13] Bartels M,Wei H. Rule-Based Improvement of Maximum Likelihood Classified LIDAR Data Fused with Co-registered Bands // Proc of the Annual Conference of the Remote Sensing and Photogrammetry Society. Cambridge,UK,2006. DOI:10.1.1.66.3166 [14] Cao Y,Wei H,Zhao H J. Optimization Algorithms in FMRF Model-Based Segmentation for LIDAR Data and Co-registered Bands // Proc of the 5th IAPR Workshop on Pattern Recognition Remote Sensing. Tampa,USA,2008: 1-4 [15] Szeliskl R,Zabih R,Scharstein D,et al. A Comparative Study of Energy Minimization Methods for Markov Random Fields with Smoothness-Based Priors. IEEE Trans on Pattern Analysis and Machine Intelligence,2008,30(6): 1068-1080 [16] Farag A A,Mohamed R M,El-Baz A. An Unified Framework for MAP Estimation in Remote Sensing Image Segmentation. IEEE Trans on Geoscience and Remote Sensing,2005,43(7): 1617-1634 [17] Platt J C. Probabilistic Outputs for Support Vector Machines and Comparison to Regularized Likelihood Methods // Smola A J,Bartlett P L,Sch lkopf B,et al.,eds. Advances in Large Margin Classifiers. Cambridge,USA: MIT Press,1999 [18] Lin H T,Lin C J,Weng R C. A Note on Platt′s Probabilistic Outputs for Support Vector Machines. Machine Learning,2007,68(3): 267-276 [19] Gao J B,Gunn S R,Harris C J. Mean Field Method for the Support Vector Machine Regression. Neurocomputing,2003,50: 391-405 [20] Wu T F,Lin C J,Weng R C. Probability Estimates for Multi-class Classification by Pairwise Coupling. Journal of Machine Learning Research,2004,5: 975-1005 [21] Bovolo F,Bruzzone L. A Context-Sensitive Technique Based on Support Vector Machines for Image Classification // Proc of the 1st International Conference on Pattern Recognition and Machine Intelligence. Kolkata,India,2005: 260-265 [22] Tarabalkay Y,Chanussot J,Benedlktsson J A. Segmentation and Classification of Hyperspectral Images Using Minimum Spanning Forest Grown from Automatically Selected Markers. IEEE Trans on Systems,Man,and Cybernetics-Part B,2010,40(5): 1267-1279 [23] Boykov Y,Veksler O,Zabih R. Fast Approximate Energy Minimization via Graph Cuts. IEEE Trans on Pattern Analysis and Machine Intelligence,2001,23(11): 1222-1239 [24] Tappen M F,Freeman W T. Comparison of Graph Cuts with Belief Propagation for Stereo Using Identical MRF Parameters // Proc of the 9th IEEE International Conference on Computer Vision. Nice,France,2003: 900-907 [25] Kolmogorov V,Rother C. Comparison of Energy Minimization Algorithms for Highly Connected Graphs // Proc of the 9th European Conference on Computer Vision. Graz,Austria,2006: 1-15 [26] Felzenszwalb P F,Huttenlocher D P. Efficient Belief Propagation for Early Vision. International Journal of Computer Vision,2006,70(1): 41-54 [27] Sun J,Zheng N N,Shum H Y. Stereo Matching Using Belief Propagation. IEEE Trans on Pattern Analysis and Machine Intelligence,2003,25(7): 787-800
2014, Vol. 27 
