基于三维分割的太阳光球亮点识别及跟踪

Identifying and Tracking Solar Photospheric Bright Points Based on Three-dimensional Segmentation Technology

光球亮点是存在于米粒暗径中小尺度和短寿命的增亮现象.提出了一个在三维时空立方体中对亮点进行识别和跟踪的方法.与普遍采用的"先识别后跟踪"策略不同,该算法采用"边识别边跟踪"的方法.该方法能够在亮点强度较弱的阶段识别出亮点,能够避免拉普拉斯形态学膨胀(LMD)算法出现的亮点演化不连续的情况,从而完整地识别到一个亮点的演化过程.通过对Hinode/SOT(Solar Optical Telescope)的G-band的一组宁静区数据的统计表明,孤立亮点的平均生命期长达3 min,最长的达到27 min,这意味着光球亮点的生命期要比以前LMD算法统计得到的更长.其强度比的均值为平均光球强度的1.02倍,比LMD算法统计的更弱,而且在生命期内出现2～3 min的周期振荡现象.

Photospheric bright points （PBPs） are tiny and short-lived phenomena which can be seen within dark inter-granular lanes. In this paper, we develop a new method to identify and track the PBPs in the three-dimensional data cube. Different from the previous way such as Detection-Before-Tracking, this method is based on the Tracking-While-Detection. Using this method, the whole lifetime of a PBP can be accurately measured while this PBP is possibly separated into several with Laplacian and morphological dilation （LMD） method due to its weak intensity sometimes. With consideration of the G-band PBPs observed by Hinode/SOT （Solar Optical Telescope） for more than two hours, we find that the isolated PBPs have an average lifetime of 3 minutes, and the longest one is up to 27 minutes, which are greater than the values detected by the previous LMD method. Furthermore, we also find that the mean intensity of PBPs is 1.02 times of the mean photospheric intensity, which is less than the values detected by LMD method, and the intensity of PBPs presents a period of oscillation with 2-3 minutes during the whole lifetime.