光电工程, 2016, 43 (4): 33, 网络出版: 2016-05-11  

半监督复合核图聚类在高光谱图像中的应用

Semi-supervised Graph Clustering with Composite Kernel and Its Application in Hyperspectral Image
作者单位
重庆大学光电技术与系统教育部重点实验室,重庆 400030
摘要
针对图的半监督聚类算法 (Semi-Supervised Graph-Based Clustering, SSGC)中出现的对先验信息利用不充分、不足以应对数据异构、计算耗时大等问题,本文提出一种基于半监督复合核的图聚类算法,并应用于高光谱图像。该算法首先通过引入半监督学习方法对径向基函数 (Radial Basis Function, RBF)进行了改进,以充分利用少量的标记样本和无标记样本;其次将 RBF核与光谱角核进行融合,构造复合核权重矩阵。在权重矩阵的构造过程中, K-近邻方法的引入也简化了计算过程。在 Indian Pine和 Botswana高光谱数据集上的实验结果表明,相对于 SSGC算法,本文算法不仅实现了更高的分类正确率,其总体分类精度提升 1%~4%,而且有效提升了运算速度。
Abstract
A semi-supervised graph-based clustering method is presented with composite kernel for the hyperspectral images, mainly to solve the problems existed in an algorithm called Semi-Supervised Graph-Based Clustering (SSGC) and improve its performance. As for the realization, it firstly reforms the Radial Basis Function (RBF) by adopting semi-supervised approach, to exploit the wealth of unlabeled samples in the image. Then, it incorporates the spectral angle kernel with RBF kernel, and constructs a composite kernel. At last, the use of K-Nearest Neighbor (KNN) method while constructing the weight matrix has greatly simplified the calculation. Experimental result in Indian Pine and Botswana hyperspectral data demonstrates that this algorithm can not only get higher classification accuracy (1%~4% higher than SSGC, 10%~20% higher than K-means and Fuzzy C-Means (FCM), but effectively improve operation speed compared with SSGC.
参考文献

[1] 李吉明,贾森,彭艳斌. 基于光谱特征和纹理特征协同学习的高光谱图像数据分类 [J].光电工程, 2012,39(11):88-94.

    LI Jiming,JIA Sen,PENG Yanbin. Hyperspectral data classification with spectral and texture features by co-training algorithm[J]. Opto-Electronic Engineering,2012,39(11):88-94.

[2] 赵春晖,齐滨. 基于模糊核加权 C-均值聚类的高光谱图像分类 [J].仪器仪表学报, 2012,33(9):2016-2021.

    ZHAO Chunhui,QI Bin. Hyperspectral image classification based on fuzzy kernel weighted c-means clustering [J]. Chinese Journal of Scientific Instrument,2012,33(9):2016-2021.

[3] 李巧兰. 基于约束的半监督聚类的图像分割算法研究 [D].西安:西安电子科技大学, 2014:1-2.

    LI Qiaolan. Semi-supervised clustering based on constraints for image segmentation [D]. Xi′an:Xidian University,2014:1-2.

[4] 冷明伟. 主动半监督聚类及其在社团检测中的应用研究 [D].兰州:兰州大学, 2014:14-23.

    LENG Mingwei. Research of active semi-supervised clustering and its application in community detection [D]. Lanzhou: Lanzhou University,2014:14-23.

[5] 陈小冬,尹学松,林焕祥. 基于判别分析的半监督聚类方法 [J].计算机工程与应用, 2010,46(6):139-143.

    CHEN Xiaodong,YIN Xuesong,LIN Huanxiang. Semi-supervised clustering approach with discriminant analysis [J]. Computer Engineering and Applications,2010,46(6):139-143.

[6] Camps-Valls Gustavo,Marsheva Tatyana V. Bandos,ZHOU Dengyong. Semi-supervised graph-based hyperspectral image classification [J]. IEEE Transactions on Geoscience and Remote Sensing(S0196-2892),2007,45(10):3044-3054.

[7] 李亚娥. 基于图的半监督分类算法研究 [D].西安:陕西师范大学, 2012:15-18. LI Ya′e. Research on graph-based semi-supervised classification algorithm [D]. Xi'an:Shanxi Normal University,2012:15-18.

[8] Papadopoulos Dimitris F,Simos Theodore E. A modified runge-kutta-nystr.m method by using phase lag properties for the numerical solution of orbital problems [J]. Applied Mathematics & Information Sciences(S1935-0090),2013,7(2):433-437.

[9] 谢娟英,郭文娟,谢维信,等. 基于样本空间分布密度的初始聚类中心优化 K-均值算法 [J].计算机应用研究, 2012, 29(3):888-892.

    XIE Juanying,GUO Wenjuan,XIE Weixin,et al. K-means clustering algorithm based on optimal initial centers related to pattern distribution of samples in space [J]. Applications Research of Computers,2012,29(3):888-892.

[10] 王峰,籍锦程,聂百胜. K-均值聚类模糊逻辑数据融合改进算法研究 [J].中北大学学报:自然科学版, 2014,35(6): 699-703.

    WANG Feng,JI Jincheng,NIE Baisheng. An improved fusion method of fuzzy logic based on k-mean clustering [J]. Journal of North University of China:Natural Science Edition,2014,35(6):699-703.

[11] 张慧哲,王坚. 基于初始聚类中心选取的改进 FCM聚类算法 [J].计算机科学, 2009,36(6):206-209.

    ZHANG Huizhe,WANG Jian. Improved fuzzy c-means clustering algorithm based on selecting initial clustering centers [J]. Computer Science,2009,36(6):206-209.

[12] Kannan S R,Ramathilagam S. Effective fuzzy c-means clustering algorithms for data clustering problems [J]. Expert Systems with Applications(S0957-4174),2012,39(7):6292–6300.

[13] ZHAO Weizhong,HE Qing,MA Huifang. Effective semi-supervised document clustering via active learning with instance-level constraints [J]. Knowledge & Information Systems(S0219-1377),2012,30(3):569-587.

[14] Ozer Sedat,CHEN Chi H. A set of new chebyshev kernel function for support vector machine pattern classification [J]. Pattern Recognition(S0031-3203),2011,44(7):1435-1447.

[15] ZHANG Rui,WANG Wenjian. Facilitating the application of support vector machine by using a new kernel [J]. Export Systems with Applications(S0957-4174),2011,38(11):14225-14230.

李志敏, 郝盼超, 黄鸿, 黄文. 半监督复合核图聚类在高光谱图像中的应用[J]. 光电工程, 2016, 43(4): 33. LI Zhimin, HAO Panchao, HUANG Hong, HUANG Wen. Semi-supervised Graph Clustering with Composite Kernel and Its Application in Hyperspectral Image[J]. Opto-Electronic Engineering, 2016, 43(4): 33.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!