基于DBSCAN聚类的毕星团成员星识别方法

Identification Method of Member Stars of Hyades Based on DBSCAN Cluster

毕星团成员星的判定始终是天文学中一项重要的任务。文章根据依巴谷卫星的观测数据来判别毕星团的成员星。在此问题中,根据毕星团中心密度大,恒星数量多的特点,采用DBSCAN密度聚类算法来判断恒星是否属于毕星团成员星,由于三维聚类计算消耗较大,于是首先利用二维投影替代三维图像。考虑到二维可能出现判断重叠的现象,又对其他变量进行聚类对结果进行修正得到了最终的成员星组成和赫罗图,结果显示模型应用性和准确性能够满足要求。

It is always an important task in astronomy to determine the member stars of Hyades.Based on the observation data of Hipparcos,this paper discriminates the member stars of Hyades.In this problem,according to the characteristics of high density and large number of stars in the center of Hyades,the DBSCAN density clustering algorithm is adopted to judge whether the stars belong to the member stars of Hyades.Due to the high consumption of 3D clustering calculation,the two-dimensional projection is first used to replace 3D image.Considering that judgment overlap may occur in two dimensions,the final member star composition and Hertzsprung-Russell diagram are obtained by clustering other variables and revising the results.The results show that the applicability and accuracy of the model can meet the requirements.