Classifying G-Band Bright Points Based on Clustering Analysis

Clustering algorithm can be used for exploring deep information from data base and outlining the characteristics of each class. In this paper, some typical clustering algorithms were reviewed. Then we focused on classifying G-band bright points (GBPs) in high resolution solar photospheric images. It is found that K-means is suitable for detecting the noise points and DBSCAN algorithm is good at extracting the boundary points.