Study On The Classification And Physical Characteristic Of Day-side Aurora

Aurora,caused by the collision of solar wind and magnetosphere's particles,is a natural phenomenon.Aurora morphology is influenced by the complex space physical environment.During the study of aurora,we have a direct way to monitor the coupling of solar wind and magnetosphere.Early research has shown that the different types of aurora are correlated with specific magnetospheric regimes and dynamic activities.Therefore,the research of aurora classification helps the analysis of the physical mechanism of each auroral type.In addition to the changes of aurora morphology,there will be some special events during the occurrence of aurora,which will also reflect some special physical process.So it is significant to classify the aurora morphology and analyze the related physical characteristics.In order to classify aurora image and figure out the relationship between aurora morphology and the space environment,an aurora image classification method based on GMM-HMM is proposed.Firstly,the GMM-HMM model is used to classify the aurora images by the static features extracted from aurora images.Furthermore,in order to analyze the influence of the different physical parameters on aurora morphology,the static features and the corresponding physical parameters are applied to classify the aurora.The aurora images not only can be classified by our proposed method,but also the relationship between aurora morphology and the physical parameters can be concluded.The aurora phenomenon is a dynamically evolving process.The aurora sequence analysis is more reasonable and desirable by utilizing temporal-spatial information instead of static feature only.In this paper,an aurora sequence classification method based on viscous fluid motion model is proposed.Firstly,the aurora particles are taken as the viscous fluid particles.Secondly,the motion feature of aurora is extracted by the viscous fluid motion model.Then the aurora sequence is classified by combining the temporal and spatial information of aurora.The experimental results demonstrate that the motion feature of aurora is extracted effectively and a high classification accuracy is obtained by the proposed method.The aurora spiral is a special event in the field of space physics,especially the aurora vortex,which reflects some special physical process.In this paper,we present an automatic method to detect the aurora spirals by using the motion information of aurora spiral from aurora sequences.Then the detected aurora spirals are divided into 4 types,the aurora arc vortex,the aurora radial vortex,the aurora arc with twisted structure and the aurora radial with twisted structure manually.Finally,the physical characteristics of these aurora spirals are analyzed by using the physical parameters from OMNI,Cluster and THEMIS satellites.The experimental results demonstrate that the aurora spiral can be detected by our proposed method.By analyzing the satellite data,we can conclude that the space environment will change with the aurora spiral event.