Study Of Auroral Oval Segmentation In UVI Images

Aurora,caused by the collision of solar wind and magnetospheric particles,are a natural phenomenon.Aurora mainly appeared in the ring region on the north and south poles of the Earth.This light ring region is called aurora oval which can be clearly observed from the UVI images which are obtained by the Polar satellite's ultraviolet imager.The boundary of auroral oval is closely related to the coupling processes of the solar wind and the Earth's magnetic field.Finding the boundary of the auroral oval accurately is of great benefit to explore the activity law of interplanetary magnetic field and geomagnetic field.Therefore,it is great significance to segment the auroral oval from the UVI images and analyze the relationship between the boundary of auroral oval and the physical parameters according to the segmentation results.To address the problems that the existing auroral oval segmentation methods are low in accuracy and sensitive to the parameters,we propose a boundary constraint factor embedded localizing active contour model for aurora oval segmentation.In this method,an active contour energy function which added the boundary constraint factor and the local information is constructed.Then the energy function is optimized using Graph cuts method.There is only one parameter in this method.And the experiment demonstrates that the parameter has little effect on the segmentation results.So,it is not necessary to adjust the parameter in our method,which means that our method can achieve the automatic auroral oval segmentation.In addition,the experimental results demonstrate that the segmentation accuracy of the proposed method is higher.We present a novel active contour model based on measure of medium truth degree to further improve the accuracy of auroral oval segmentation.In this paper,a novel active contour energy function based on measure of medium truth degree are constructed.This algorithm is robust to noise,and can obtain more accurate boundary.In addition,the experimental results show that the proposed algorithm also has outstanding performance on natural images,artificial images,and medical images,which demonstrates that the proposed method has outstanding universality.For UVI sequences segmentation,the static information of the UVI images are only considered and the motion characteristics of the auroral particles are always neglected in the existing methods.So,a new method for auroral oval sequences segmentation based on static information and dynamic information of UVI sequences is proposed.Firstly,a method based on eye movement guiding and patches matching is used to obtain the spatial auroral oval segmentation results.Then the Lattice Boltzmann method is used to obtain the particle motion field of auroral oval.And the temporal auroral oval segmentation results can be obtained according to the particle motion field.The final segmentation results can be obtained by combining the spatial auroral oval segmentation results and the temporal auroral oval segmentation results.Then we analyze the relationship between the auroral oval boundary and the physical parameters.Firstly,the dynamic information of auroral oval is added to the BP-Ada Boost model to predict the auroral oval boundary.The statistical results of the relationship between the auroral oval boundary and the physical parameters show that the geomagnetic index,the interplanetary magnetic field and solar wind have different effects on the boundaries of auroral oval.