A Modified FCM Classifier Constrained by Conditional Random Field Model for Remote Sensing Imagery

doi:10.11947/j.AGCS.2016.20150621

Abstract

Abstract: Remote sensing imagery has abundant spatial correlation information, but traditional pixel-based clustering algorithms don't take the spatial information into account, therefore the results are often not good. To this issue, a modified FCM classifier constrained by conditional random field model is proposed. Adjacent pixels' priori classified information will have a constraint on the classification of the center pixel, thus extracting spatial correlation information. Spectral information and spatial correlation information are considered at the same time when clustering based on second order conditional random field. What's more, the global optimal inference of pixel's classified posterior probability can be get using loopy belief propagation. The experiment shows that the proposed algorithm can effectively maintain the shape feature of the object, and the classification accuracy is higher than traditional algorithms.

Key words: clustering for remote sensing imagery, fuzzy C-means, conditional random field, spatial correlation information

CLC Number:

P237

WANG Shaoyu, JIAO Hongzan, ZHONG Yanfei. A Modified FCM Classifier Constrained by Conditional Random Field Model for Remote Sensing Imagery[J]. Acta Geodaetica et Cartographica Sinica, 2016, 45(12): 1441-1447.

References

[1] TAŞDEMIR K, MERÉNYI E. A Validity Index for Prototype-based Clustering of Data Sets with Complex Cluster Structures[J]. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 2011, 41(4):1039-1053.
[2] JAIN A K. Data Clustering:50 Years beyond K-means[J]. Pattern Recognition Letters, 2010, 31(8):651-666.
[3] MAULIK U, SAHA I. Modified Differential Evolution Based Fuzzy Clustering for Pixel Classification in Remote Sensing Imagery[J]. Pattern Recognition, 2009, 42(9):2135-2149.
[4] LLOYD S. Least Squares Quantization in PCM[J]. IEEE Transactions on Information Theory, 1982, 28(2):129-137.
[5] BALL G H, HALL D J. ISODATA, A Novel Method of Data Analysis and Pattern Classification[R]. Technical Report AD0699616. Menlo Park, CA:Stanford Research Institute, 1965.
[6] 李雷, 罗红旗, 丁亚丽. 自适应约束模糊C均值聚类算法[J]. 模糊系统与数学, 2010, 24(5):126-130. LI Lei, LUO Hongqi, DING Yali. Adaptive Constrained FCM Clustering Algorithm[J]. Fuzzy Systems and Mathematics, 2010, 24(5):126-130.
[7] 耿连英, 贺兴时, 贺飞跃. 基于模糊分类的MRF图像恢复方法[J]. 计算机工程, 2011, 37(10):198-199, 221. GENG Lianying, HE Xingshi, HE Feiyue. MRF Image Restoration Method Based on Fuzzy Classification[J]. Computer Engineering, 2011, 37(10):198-199, 221.
[8] ZHONG Yanfei, MA Ailong, ZHANG Liangpei. An Adaptive Memetic Fuzzy Clustering Algorithm with Spatial Information for Remote Sensing Imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(4):1235-1248.
[9] BANDYOPADHYAY S, MAULIK U, MUKHOPADHYAY A. Multiobjective Genetic Clustering for Pixel Classification in Remote Sensing Imagery[J]. IEEE Transactions on Geoscience and Remote Sensing, 2007, 45(5):1506-1511.
[10] GHOSH A, MISHRA N S, GHOSH S. Fuzzy Clustering Algorithms for Unsupervised Change Detection in Remote Sensing Images[J]. Information Sciences, 2011, 181(4):699-715.
[11] NIAZMARDI S, HOMAYOUNI S, SAFARI A. An Improved FCM Algorithm Based on the SVDD for Unsupervised Hyperspectral Data Classification[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2013, 6(2):831-839.
[12] AHMED M N, YAMANY S M, MOHAMED N, et al. A Modified Fuzzy C-means Algorithm for Bias Field Estimation and Segmentation of MRI Data[J]. IEEE Transactions on Medical Imaging, 2002, 21(3):193-199.
[13] 刘启亮, 邓敏, 石岩, 等. 一种基于多约束的空间聚类方法[J]. 测绘学报, 2011, 40(4):509-516. LIU Qiliang, DENG Min, SHI Yan, et al. A Novel Spatial Clustering Method Based on Multi-constraints[J]. Acta Geodaetica et Cartographica Sinica, 2011, 40(4):509-516.
[14] HAMASUNA Y, ENDO Y, MIYAMOTO S. On Mahalanobis Distance Based Fuzzy C-means Clustering for Uncertain Data Using Penalty Vector Regularization[C]//Proceedings of 2011 IEEE International Conference on Fuzzy Systems. Taipei, Taiwan, China:IEEE, 2011:810-815.
[15] 杨红磊, 彭军还. 基于马尔可夫随机场的模糊C-均值遥感影像分类[J]. 测绘学报, 2012, 41(2):213-218. YANG Honglei, PENG Junhuan. Remote Sensing Classification Based on Markov Random Field and Fuzzy C-Means Clustering[J]. Acta Geodaetica et Cartographica Sinica, 2012, 41(2):213-218.
[16] 张路, 廖明生. 一种顾及上下文的遥感影像模糊聚类[J]. 遥感学报, 2006, 10(1):58-65. ZHANG Lu, LIAO Mingsheng. Contextual Fuzzy Clustering of Remote Sensing Imagery[J]. Journal of Remote Sensing, 2006, 10(1):58-65.
[17] 姜文志, 顾佼佼, 胡文萱, 等. 条件随机场模型的应用研究及改进[J]. 计算机与现代化, 2011(11):55-58. JIANG Wenzhi, GU Jiaojiao, HU Wenxuan, et al. Research on Applications of Conditional Random Fields and Its Improvement[J]. Jisuanji Yu Xiandaihua, 2011(11):55-58.
[18] 李祖传, 马建文, 张睿, 等. 利用SVM-CRF进行高光谱遥感数据分类[J]. 武汉大学学报(信息科学版), 2011, 36(3):306-310. LI Zuchuan, MA Jianwen, ZHANG Rui, et al. Classifying Hyperspectral Data Using Support Vector Machine Conditional Random Field[J]. Geomatics and Information Science of Wuhan University, 2011, 36(3):306-310.
[19] ZHONG Yanfei, LIN Xuemei, ZHANG Liangpei. A Support Vector Conditional Random Fields Classifier with a Mahalanobis Distance Boundary Constraint for High Spatial Resolution Remote Sensing Imagery[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7(4):1314-1330.
[20] WANG Yang, RAJAPAKSE J C. Contextual Modeling of Functional MR Images with Conditional Random Fields[J]. IEEE Transactions on Medical Imaging, 2006, 25(6):804-812.
[21] SYMONS C T, SAMATOVA N F, KRISHNAMURTHY R, et al. Multi-criterion Active Learning in Conditional Random Fields[C]//Proceedings of the 18th IEEE International Conference on Tools with Artificial Intelligence. Arlington, VA:IEEE, 2006:323-331.
[22] NOMOTO T. Discriminative Sentence Compression with Conditional Random Fields[J]. Information Processing & Management, 2007, 43(6):1571-1587.
[23] ZHANG Guangyun, JIA Xiuping. Simplified Conditional Random Fields with Class Boundary Constraint for Spectral-Spatial Based Remote Sensing Image Classification[J]. IEEE Geoscience and Remote Sensing Letters, 2012, 9(5):856-860.