LAMOST天体光谱科学参数的自动测量及特殊天体目标的自动搜索算法研究

Guoshoujing Telescope (the Large Sky Area Multi-Object Fiber Spectroscopic Telescope, LAMOST) is a National Major Scientific Project built by the Chinese Academy of Sciences. With the survey plan being carried out in an orderly manner, LAMOST has obtained more than 7,000,000 spectra data which include pilot survey data and data release 1-4 of regular survey. However, massive spectral data needs efficient automatic processing algorithm. This project is supposed to focus on the spectral data of LAMOST in automatic processing and analysis, applying computer technology and methods to conduct the research on the automatic processing of massive astronomical spectra. The main research contents of this project include the automatic measurement of the scientific parameters about the celestial spectrum (redshift, stellar atmospheric parameters), and the search for rare celestial object candidates from the LAMOST spectral data. In this project, we will apply mathematical theory and computer technology to achieve the redshift measurement of extragalactic celestial objects, measurement of stellar atmospheric parameters and data mining of rare celestial object automatically. The research contents are as follows: (1) The research on automatic redshift measurement algorithms for extragalactic celestial spectrum from the LAMSOT; (2) The research on automatic algorithms of stellar atmospheric parameters measurement for the Stellar spectra in the Milky way from the LAMOST; (3) The research on auto-searching algorithms about the rare celestial objects in LAMOST spectral data; (4) Design and implementation of the software system for automatic scientific parameters measurement and the search for special celestial objects in LAMOST massive spectral data.

郭守敬望远镜（又称LAMSOT）是我国天文观测设备中居于国际领先的大科学装置，至今已释放包括先导巡天、正式巡天1-4期数据在内的700多万条光谱，如此海量的天体光谱需要高效准确的自动处理方法来为天文学家提供科学数据。本课题基于郭守敬望远镜的光谱自动处理分析需求，结合天文学背景，拟利用数学理论和计算机技术与方法，开展包括天体光谱科学参数（红移、恒星大气参数）的自动测量和稀有天体目标候选体的搜寻方面算法的研究。具体研究内容包括：1、针对LAMOST的河外光谱，研究高效稳定的红移自动测量算法；2、研究LAMOST所获恒星光谱的大气参数自动测量算法；3、研究特殊天体目标的自动搜寻算法；4、设计实现针对LAMOST海量光谱的科学参数自动测量和特殊天体目标搜寻的软件系统。

在课题研究过程中，课题组严格执行了合同研究计划，完成了合同规定的研究内容，在天体光谱数据分析与处理等天文信息处理方面取得了一系列具有重要价值的研究成果。本课题共发表学术论文21篇，其中 SCI检索论文17篇，国际会议EI论文4篇。本课题基于郭守敬望远镜的光谱自动处理分析需求，结合天文学背景，利用数学理论和计算机技术与方法，开展了包括天体光谱自动分类、科学参数的自动测量以及参数优化算法方面研究。具体研究内容和成果体现在以下三个方面：1、研究提出了三种SVM的改进算法和一种改进的基于曼哈顿距离密度聚类的算法来实现天体光谱的自动分类；2、研究提出了核主成分弹性回归、核岭回归以及极端随机树等三种恒星参数自动测量算法；3、研究提出了改进粒子群算法等8种参数优化算法来实现机器学习算法中多目标优化的参数调优问题。