基于FastICA和高斯滤波的星系团检测算法  

A cluster detection method based on FastICA and gauss filter

作　　者：吴中耀[1] 马志贤[1] 朱杰[1] WU Zhong-yao MA Zhi-xian ZHU Jie(D epartm ent o f Electronic E ngineering,Shanghai Jiaotong U niversity,Shanghai 200240,C hin)

机构地区：[1]上海交通大学电子工程系,上海200240

出　　处：《信息技术》2016年第11期26-28,共3页Information Technology

基　　金：国家自然科学基金项目资助(11433002;61271349;613-71147)

摘　　要：星系团作为宇宙中大尺度结构的基本单元,是天文观测中需要提取的重要成分。由于宇宙其他强宇宙前景辐射的混叠,使得提取星系团辐射成为一个难题。文中使用基于FastICA和高斯滤波结合的算法进行星系团检测,首先使用FastICA对模拟的低频射电的叠加宇宙辐射进行盲源分离,然后对分离后的带有其他辐射残留的星系团成分进行高斯滤波和峰值检测,达到星系团的识别的目的。实验表明本方法的识别率可以达到75.7%,为进一步提取星系团的成分奠定了良好基础。Cluster is an important composition which is observed in astronomy as the basic element of big scale construction in universe. It has been a challenge for extracting radiation from cluster because of the mix of other strong foregrounds. In this paper,a new method,which is based on FastICA and Gauss filter,is implemented to separate the superimposed cosmic radiation in low frequency radio first,then it convolves the separated cluster contaminated with other radiation with Gauss filter and peak detection,to reach the purpose of cluster detection. The detection rate is 75. 7%,and this is a good foundation for further extracting cluster component.

关 键 词：星系团检测 FASTICA 高斯滤波 低频射电 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]