Aurora Sequences Classification and Aurora Events Detection Based on Hidden Conditional Random Fields

摘要：

　　The dynamically evolving process of aurora is closely related to complex and energetic plasma processes of the outer magnetosphere,so aurora image sequences often have complex underlying structures.In this paper,we present a novel aurora sequences classification and aurora events detection method based hidden conditional random fields(HCRF)employing spatial texture features.Firstly,divided uniform local binary patterns(uLBP)are extracted as the spatial texture features; then HCRF model is built for the spatial texture features of aurora sequences; at last,the model is applied in automatic classification and detection for four primary categories of dayside auroral observations.The supervised classification results on labeled data demonstrate the effectiveness of our method.The occurrence distributions of four categories from automatic detection confirm the multiple-wavelength intensity distribution of dayside aurora,and further illustrate the validity of our method.

关键词： Aurora sequence classification Aurora events detection Hidden conditional random fields (HCRF) Aurora morphology

作者: Baibai Xu Changhong Chen Zongliang Gan Bin Liu

作者单位: Jiangsu Provincial Key Lab of Image Processing and Image Communication,Nanjing University of Posts and Telecommunications,Nanjing 210003,China

会议类型: 国际会议

会议名称: 第七届全国模式识别学术会议(The 7th Chinese Conference on Pattern Recognition,CCPR2016)

会议地点: 成都

会议语种:英文

页码: 404-415

在线出版日期: 2016-11-03（万方平台首次上网日期，不代表论文的发表时间）