基于SIFT特征检测和密度峰值聚类的太阳活动区自动检测算法研究

太阳活动区是太阳大气中产生各种活动现象的区域,精确地检测和识别太阳活动区对理解太阳磁场的形成机制具有极为重要的科学意义.根据太阳活动区结构较为复杂的特点,基于尺度不变特征变换(ScaleInvariant Feature Transform, SIFT)和密度峰值聚类(Clustering by Fast Search and Find of Density Peaks,DPC)算法的优越性,提出了一种太阳活动区的自动检测和识别方法.首先,对太阳动力学天文台(Solar Dynamics Observatory, SDO)日震和磁场成像仪(Helioseismic and Magnetic Imager, HMI)的纵向磁图进行对比度增强;然后采用SIFT方法提取出全日面磁图中的特征点;最后利用DPC算法将特征点进行聚类,从而自动检测和识别出太阳活动区.研究结果表明, SIFT和DPC算法相结合的方法可以在不需要人工交互的情况下准确地自动检测出太阳活动区.

An Automatic Detection of Solar Active Regions Based on Scale-Invariant Feature Transform and Clustering by Fast Search and Find of Density Peaks

The solar active regions are the sites of various activities taking place in the solar atmosphere. Accurate detection and identification of the solar active regions are of great scientific significance to understand the formation mechanism of the solar magnetic field. In this paper, we propose an automatic detection and recognition method for solar active regions based on the advantages of Scale-Invariant Feature Transform (SIFT) and Clustering by Fast Search and Find of Density Peaks (DPC). Firstly, contrast enhancement is used in the longitudinal magnetic image of Helioseismic and Magnetic Imager (HMI) of Solar Dynamics Observatory (SDO). Then, the feature points are extracted by SIFT. Finally, the feature points are clustered by fast search and find of density peaks so as to automatically detect and identify the solar active regions. The results show that the combination of SIFT and DPC can accurately identify the solar active regions without human-computer interaction.