Dayside corona aurora classification based on X-grey level aura matrices and feature selection

Abstract

Dayside corona aurora is the main form of aurora at magnetic noon, which is generated by the dynamics process of the interaction of the sun and earth's magnetosphere. Hence the study of dayside corona aurora is of great importance to the analysis of ionosphere and its dynamic feature. This paper proposes a novel aurora texture extraction method based static image classification of dayside aurora, in which X-grey level aura matrices are designed to extract the features of the original aurora images. Besides, a dayside aurora classification algorithm for dayside aurora based on feature selection is proposed to handle the large quantities of aurora samples. To eliminate the effect of noise and solve the problem of high feature dimensionality, the ReliefF is adopted to select the effective feature vector. For classification, texture models are learned by using the support vector machine, then a given texture of aurora image can be classified into one of the pre-learned classes. All of the aurora images in the experiment are obtained from the all-sky aurora image system in the Chinese Arctic Yellow River Station. The experimental results illustrate the high effectiveness of the proposed dayside aurora classification algorithm.

Keywords:

• dayside corona aurora
• X-GLAM
• SVM
• ReliefF
• texture classification
• feature selection

2010 AMS Subject Classifications :

• 62H30
• 68T10
• 68U10

Acknowledgements

This research was supported by the National Natural Science Foundation of China (Nos. 41031064, 60902082), the Ocean Public Welfare Scientific Research Project, State Oceanic Administration of China (No. 201005017), the Fundamental Research Funds for the Central Universities (No. JY10000902016), Natural Science Basic Research Plan in Shaanxi Province of China (No. 2011JQ8019) and the PhD Programs Foundation of Ministry of Education of China (No. 20090203110002).