Aurora Image Classification Based On Global Morphological Features

The aurora is a natural phenomenon caused by the collision of high energy charged particles moving along the magnetic field and the polar upper atmosphere.The aurora phenomenon reflects the dynamic process of solar wind's high energy particles entering into the polar region,which is one of the most important physical phenomena in the solar energy coupling system.Different shape of the aurora have different physical formation process and physical features.It is helpful to study the mechanism of aurora occurrence and the relationship between auroral and magnetospheric dynamics.Since the 50s of last century,scientists around the world have been working on the aurora research.However,the understanding of the detailed physical processes and the formation mechanism behind different auroral forms is still not fully understood,the research of the aurora classification has been a hot research topic in this field.In recent years,more and more researchers applied computer image processing and pattern recognition technology into space physics is used to achieve the characterization and analysis of image data,this method has been applied to the field of aurora classification and has been achieved remarkable results.Some of the existing classification methods of aurora images achieved the classification of aurora images based on the outline of the aurora images.Some of the methods achieved classification of aurora images based on the subspace,while others pay more attention to the texture information of aurora images.In recent years,some researchers connect human sensory system and psychology with theaurora image classification.The majority of the methods are based on some local invariant descriptors to extract the local features of the aurora images to realize the automatic classification of the aurora images.These methods have obtained good classification results,but the local features are sensitive to noise,variances of the position and orientation.It is difficult to meet the requirements about robust feature descriptor in actual application.To solve this problem,this paper proposed a novel aurora image classification method based on the global feature descriptors.In the proposed method,the aurora images after preprocessing are projected to radon domain via radon transform,and then,the variances of columns are determined,the rotation invariant features are obtained via circular shift operation on the variance sequence to let the maximum value in the first place,which completes the rotation normalization of the variance sequence.A nearest neighbor classifier based on Euclidean distance is used for classification.In order to verify the proposed method excellent characterization ability,some comparison experiments are carried out to compare the proposed method with other existing classification methods based on global feature and local features.Experimental results show that the proposed approach yields a better performance in terms of the correct classification percentage compared with the other existing aurora image classification methods.It is also shown that the proposed approach yields observably low computational cost and relatively high robustness to noise,the variations of orientation and position of aurora images.