The Research And Implementation Of Solar Filament Detection Method And Chirality Recognition Method

The research of solar filament has important significance on analysis of the geophysical effect of solar activity and the solar magnetic field. Chirality is a property to describe solar filament, it can reflect the changes of magnetic field. The inadequate of traditional filament detection methods is that it can’t detect limbic filaments, mixed sunspots, miss smaller filaments, etc, while the new methods cost a lot of time. Spine recognition is an important part of filament chirality recognition, the detection result of existing spine recognition methods is not very accurate, and the algorithms are complicated and time-consuming. Existing chirality judgment methods exists a situation that it can’t judge chirality at sometime, and its consideration is too single. These problems need to be resolved promptly. According to the above problems, two new methods were proposed: filament detection method and chirality recognition method, chirality recognition including spine recognition and chirality judgment, the work of this paper mainly focus on the following four sections:The first section is filament detection. Filament detection is mainly combine limb darkening removing and morphological operations, region growing to extract filaments, most of the filaments are detected accurately, including limbic filaments and small filaments, finally, exclude sunspots on the basis of geometric parameters of eccentricity, the processing result is good.The second section is spine recognition. The paper is based on a algorithm of binary tree traversal seek to determine the longest path to recognize filament spine, the result of this method is more accurate and higher substitution than other methods, this algorithm is also used to identify the longest branch of barb in latter work.The third section is barbs’ angle calculation. Calculating barbs’ angle is calculates the angle between spine and horizontal direction and angle between barb and horizontal direction respectively at first, then, calculating barbs’angle on this basis. Due to the polymorphism of barbs, discussion about taking the barbs’ fitting points can be divided into five situations, finally determine all points on the longest branch of barb are the optimal fitting points.The fourth section is chirality judgement. We consider that the chirality of filament is not only related to the number of barbs, but also related to barbs’angle and length, in order to overcome the disadvantages of traditional filament chirality recognition methods, we proposed a weighted method to judge filament chirality, evaluating and comparing its’ performance with existing methods, the result proves weighted method is reliable, accurate and efficient to judge filament chirality.The innovation of this paper has three:Firstly, morphological closing operation was used to connect broken filaments and eliminate sunspot on the basis of eccentricity in filaments detection method. Secondly, based on algorithm of binary tree traversal seek to determine the longest path to recognize filament spine. Thirdly, judge chirality on the basis of weighted method. The methods proposed in the paper overcoming some defects exist on other methods, and it can provide the basis of reliable and strong support for analyzing solar activity, geophysical effect of solar activity, solar magnetic field and other scientific research.