南京大学学报(自然科学版) ›› 2016, Vol. 52 ›› Issue (1): 184193.
宋亚婷1, 韩 冰1,2*, 高新波1,
Song Yating1, Han Bing1,2*, Gao Xinbo1
摘要: 极光事件的形态与行星际磁场及太阳风等空间物理过程密切相关,通过研究极光形态可以得到大量地球磁层和太阳风活动的信息.随着海量极光的产生,如何借助计算机对极光图像序列进行自动分类成为热点.目前极光分类研究大多是基于单幅图像的特征分析,极光视频的建模和分析仍然有待深入研究.提出一种基于极光视频动态纹理建模的极光视频事件识别方法.首先对四类极光视频进性普适性动态纹理建模.该模型能充分表征极光序列帧间的重复相关性和动态变化特性.然后利用矩阵SVD分解对动态纹理模型求解,最后用模型参数间的马丁距离衡量极光序间差异性,采用最小距离分类器和SVM分类器实现四类典型形态的极光序列的自动分类识别.为进一步提高模型紧凑型,引入张量分解,建立张量动态纹理模型.相比于动态纹理模型只关注序列帧间的重复相关性,张量动态纹理模型还分析图像帧内各个部分间的重复相关性.张量动态纹理模型能从时间和空间维度上同时进行分解,提取模型参数构造极光序列特征,减少模型冗余的同时提高分类准确率.在中国北极黄河站的ASI图像数据库上进行了算法验证,实验结果表明该方法分类准确率可以达到80%以上.算法能有效识别不同形态的极光序列事件.
[1]Pedersen T R, Gerken E A. Creation of visible artificial optical emission in the aurora by high-power radio waves. Nature, 2005, 433(7025): 498-500. [2]Hu Z J, Yang H, Huang D, et al. Synoptic distribution of dayside aurora: Multiple-wavelength all-sky observation at Yellow River Station in Ny-Alesund, Svalbard. Journal of Atmospheric and Solar-terrestrial Physics, 2009, 71(89): 794-804. [3]Lorentzen D A, Moen J, Oksavik K, et al. In situmeasurement of a newly created polar cap patch. Journal of Geophysical Research ,2010, 115(A12). [4]Zhang Q H, Zhang B C, Michael L, et al. Direct observations of the evolution of polar cap ionization patches. Science, 2013, 339:1597-1600. [5]Syrjasuo M, Partamies N. Numeric image features for detection of aurora. Geoscience and Remote Sensing Letters, 2012, 9(2): 176-179. [6]王 倩,梁继民,高新波等. 基于表象特征的极光图形分类方法研究. 第12届全国日地空间物理学术研讨会论文摘要集, 2010, 72(5):498–508. [7]Gao L, Gao X B, Liang J M. Dayside corona autora detection based on sample selection and adaBoost algorithm. Journal of Image and Graphics, 2010, 15(1): 116-121. [8]Ru F, Li J, Gao X B, et al. Automatic aurora images classification algorithm based on separated texture. In: Proceedings of IEEE International Conference on Robotics and Biomimetics, 2009: 1331-1335. [9]Wang Y R, Gao X B, Fu R. Dayside corona aurora classification based on X-gray level aura matrices. In: Proceedings of the ACM International Conference on Image and Video Retrieval. ACM, 2010. 282-287. [10]Han B, Zhao X, Tao D, et al. Dayside aurora classification via BIFs-based sparse representation using manifold learning. International Journal of Computer Mathematics. Published online: 12 Nov 2013 [11]杨 曦,李 洁,韩 冰等.一种分层小波模型下的极光图像分类算法. 西安电子科技大学学报(自然科学版), 2013,40(2):18-24 . [12]韩 冰,杨 辰,高新波.融合显著信息的LDA极光图像分类.软件学报,2013,24(11):2758?2766 . [13]Yang Q. Auroral events detection and analysis based on ASI and UVI images. Doctoral Dissertation. Xi’an : Xidian University,2013. [14]韩 冰,廖 谦. 基于空时极向LBP的极光序列事件检测.软件学报, 2014,25(9):2172-2179. [15]Soatto S,Doretto G,Wu Y N. Dynamic textures. In: Proceedings of IEEE International Conference on Computer Vision, 2001, 2:439-446. [16]Avinash R, Rizwan C, Rene V. Categorizing dynamic textures using a bag of dynamical systems. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(2):342-53. [17]Mumtaz A,Coviello E,Lanckriet G R, et al. Clustering dynamic textures with the hierarchical em algorithm for modeling video. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2013,35(7):1606-21. [18]Sloven D, Renaud P, Michel M. Decomposition of dynamic textures using morphological Component Analysis. Transactions on Circuits and Systems for Video Technology, 2011, 1-14. [19]易兴辉,王国胤,胡 峰. 一种新的基于粗糙集的动态样本识别算法.南京大学学报(自然科学), 2010, 46(5):501-506. [20]于振洋,高尚兵,唐嵩涛.应用局部结构与方向张量的图像分割算法研究.南京大学学报(自然科学), 2015, 51(1):111-117. [21]Lathauwen L D, Moor B D, Vandewalle J. On the best rank-1and rank-(r1,r2,..r1) approximation of higher-order tensors. SIAM Journal on Matrix Analysis and Applications, 2000. [22]Peng Y, Meng D, Xu Z, et al. Decomposable nonlocal tensor dictionary learning for multispectral image denoising. Computer Vision and Pattern Recognition, 2014, 2949 – 2956. |
No related articles found! |
|