Fuzzy Support Vector Machine for PolSAR Image Classification

Abstract:

Fully Polarimetric Synthetic Aperture Radar (PolSAR) image classification, with the complexity for its data’s scattering mechanism and statistical property, has expected to be performed by an automatic categorization. This paper presents a supervised method called Fuzzy support vector machine (FSVM), which is a variant of the SVM algorithm to classify the PolSAR image data. In order to take advantages of PolSAR data, five scattering features (entropy, total power, three Eigenvalues of Coherent Matrix: λ1,λ2,λ3) are input as original data space of the FSVM algorithm. The feasibility of this approach is examined by the JPL/AIRSAR PolSAR data. The classification results show that the proposed FSVM method has out-performed the SVM method.

• Periodical: Advanced Materials Research (Volumes 639-640)
• Pages: 1162-1167
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.639-640.1162
• Online since: January 2013
• Authors: Hong Xia Ke, Guo Dong Liu, Guo Bing Pan
• Keywords: Feature Selection, Fuzzy Support Vector Machine, PolSAR Image Classification

References

[1] Ertin E, Potter L C, Polarimetric classification of scattering centers using M-ary Bayesian decision rules, IEEE Trans. on Aerospace and Electronic System, 36(2000) 738-749.

DOI: 10.1109/7.869492

[2] Kouskoulas Y, Ulaby F T, Pierce L E., The Bayesian hierarchical classifier(BHC) and its application to short vegetation using multifrequency polarimetric SAR, IEEE Trans. on Geoscience and Remote Sensing, 42(2004):469-477.

DOI: 10.1109/tgrs.2003.821066

[3] Hellmann M, Jager G, Kratzschmar E, Habermeyer M, Classification of full polarimetric SAR-data using artificial neural networks and fuzzy algorithms, in Proceedings of IEEE International Geosience and Remote Sensing Symposium, 1999: 1995-1997.

DOI: 10.1109/igarss.1999.775011

[4] Chen C T, Chen K S, Lee J S, The use of fully polarimetric information for the fuzzy neural classification of SAR images, IEEE Trans. on Geoscience and Remote Sensing, 41(2003) 2089-2100.

DOI: 10.1109/tgrs.2003.813494

[5] Fukuda S, Hirosawa H, A wavelet-based texture feature set applied to classification of multifrequency polarimetric SAR images, IEEE Trans. on Geoscience and Remote Sensing, 37(1999) 2282-2286.

DOI: 10.1109/36.789624

[6] Aiazzi B, Alparone L, Baronti S, Garzelli A, Land cover classification of built-up areas through enhanced fuzzy nearest-mean reclustering of textural features from X- and C-band polarimetric SAR data, in Proceedings of SPIE, 2004: 105-115.

DOI: 10.1117/12.514258

[7] Lin C, Wang S, Fuzzy support vector machines, IEEE Transaction on Neural Networks, 13(2002) 464-471.

DOI: 10.1109/72.991432

[8] Cloude, S.R. and Pottier, E., A review of target decomposition theorems in radar polarimetry, IEEE Transactions on Geoscience and Remote Sensing, 34(1996) 498-518.

DOI: 10.1109/36.485127

[9] Fang Cao, Wen Hong, Yirong Wu and Eric Pottier, An unsupervised segmentation with an adaptive number of clusters using the SPAN/H//A space and complex Wishart clustering foe fully Polarimetric SAR data analysis, IEEE Transactions on Geoscience and Remote Sensing, 45(2007):3454-3467.

DOI: 10.1109/tgrs.2007.907601

[10] Tai-Yue Wang, Huei-Ming Chiang, Fuzzy support vector machine for multi-class text categorization, Information Processing and Management, 43:914-292, 2007.

DOI: 10.1016/j.ipm.2006.09.011

[11] Rukshan Batuwita, Vasile Palade, FSVM-CIL:Fuzzy support vector machines for class imbalance learning, IEEE Transactions on Fuzzy Systems, 18(2010558-571.

DOI: 10.1109/tfuzz.2010.2042721