Abstract：Separability Index (SI) can be used to select effective classification features, but in the case of multi-dimensional features and good separability of geology, the use of separability index for feature selection can not effectively remove redundancy. Based on this, a method of feature selection and multi-layer Support Vector Machine (SVM) classification was proposed by using separability index and Sequential Backward Selection (SBS) algorithm. Firstly, the classification object and features were determined according to the SIs of all the ground objects under all the features, and then based on the classification accuracies of the objects, the SBS algorithm was used to select the features again. Secondly, the features of next ground objects were determined by the separability index of remaining objects and the SBS algorithm in turn. Finally, the multi-layer SVM was used for classification. The experimental results show that the classification accuracy of the proposed method is improved by 2% compared with the method of multi-layer SVM classification where features are selected only based on the SI, and the classification accuracy of all kinds of objects is higher than 86%, and the running time is half of the original method.
[1] 李仲森.极化雷达成像基础与应用[M].北京:电子工业出版社,2013:2.(LI Z S. Polarization Radar Imaging Foundation and Application[M]. Beijing:Publishing House of Electronics Industry, 2013:2.) [2] 王超,张红,陈曦.全极化合成孔径雷达图像处理[M].北京:科学出版社,2008:107.(WANG C, ZHANG H, CHEN X. Full Polarimetric Synthetic Aperture Radar Image Processing[M]. Beijing:Science Press, 2008:107.) [3] 吴永辉,计科峰,郁文贤.SVM全极化SAR图像分类中的特征选择[J].信号处理,2007,23(6):877-881.(WU Y H, JI K F, YU W X. Feature selection in SVM full polarimetric SAR image classification[J]. Signal Processing, 2007, 23(6):877-881.) [4] 吴永辉,计科峰,李禹,等.利用SVM的极化SAR图像特征选择与分类[J].电子与信息学报,2008,30(10):2347-2351.(WU Y H, JI K F, LI Y, et al. Characteristic selection and classification of polarimetric SAR images using SVM[J]. Journal of Electronics & Information Technology, 2008, 30(10):2347-2351.) [5] 巫兆聪,欧阳群东,李芳芳.顾及特征优化的全极化SAR图像SVM分类[J].测绘科学,2013,38(3):115-117.(WU Z C, OUYANG Q D, LI F F. SVM classification of fully polarimetric SAR images with feature optimization[J]. Surveying Science, 2013, 38(3):115-117.) [6] 代琨,于宏毅,马学刚,等.基于支持向量机的特征选择算法综述[J].信息工程大学学报,2014,15(1):85-91.(DAI K, YU H Y, MA X G, et al. Research on feature selection algorithm based on support vector machine[J]. Journal of Information Engineering University, 2014, 15(1):85-91.) [7] 袁春琦,徐佳,程圆娥,等.顾及分类器参数的全极化SAR图像特征选择与分类[J].测绘科学技术学报,2016,33(5):507-512.(YU C Q, XU J, CHENG Y E, et al. Design and classification of full polarized SAR image based on classifier parameters[J]. Journal of Surveying and Mapping Science and Technology, 2016, 33(5):507-512.) [8] 李雪薇,郭艺友,方涛.基于对象的合成孔径雷达影像极化分解方法[J].计算机应用,2014,34(5):1473-1476.(LI X W, GUO Y Y, FANG T. Based on the object-based synthetic aperture radar image polarization decomposition method[J]. Journal of Computer Applications, 2014, 34(5):1473-1476.) [9] 宋超,徐新,桂容,等.基于多层支持向量机的极化合成孔径雷达特征分析与分类[J].计算机应用,2017,37(1):244-250.(SONG C, XU X, GUI R, et al. Analysis and classification of polarized synthetic aperture radar based on multi-layer support vector machine[J]. Journal of Computer Applications, 2017, 37(1):244-250.) [10] CUMNLING I G, ZYL J J V. Feature utility in polarimetric radar image classification[C]//IGARSS'89:Proceedings of the 1989 International Canadian Symposium on Remote Sensing. Piscataway, NJ:IEEE, 2002:1841-1846. [11] MISHRA P, SINGH D. A statistical-measure-based adaptive land cover classification algorithm by efficient utilization of polarimetric SAR observables[J]. IEEE Transactions on Geoscience & Remote Sensing, 2014, 52(5):2889-2900. [12] WU F, WANG C, ZHANG H, et al. Rice crop monitoring in south China with RADARSAT-2 quad-polarization SAR data[J]. IEEE Geoscience & Remote Sensing Letters, 2011, 8(2):196-200. [13] GUPTA S, SINGH D, KUMAR S. An approach based on texture measures to classify the fully polarimetric SAR image[C]//Proceedings of the 2015 International Conference on Industrial and Information Systems. Piscataway, NJ:IEEE, 2015:1-6. [14] JAIN A, SINGH D. Decision tree approach to classify the fully polarimetric RADARSAT-2 data[C]//Proceedings of 2015 National Conference on Recent Advances in Electronics & Computer Engineering (RAECE). Piscataway, NJ:IEEE, 2015:318-323. [15] VAPNIK V N. An overview of statistical learning theory[J]. IEEE Transactions on Neural Networks, 1999, 10(5):988-99. [16] PUDIL P, NOVOVICOVÁ J, KITTLER J. Floating search methods in feature selection[J]. Pattern Recognition Letters, 1994, 15(11):1119-1125. [17] LANG F, YANG J, LI D, et al. Polarimetric SAR image segmentation using statistical region merging[J]. IEEE Geoscience & Remote Sensing Letters, 2014, 11(2):509-513.
2018, Vol. 38 
