基于模板匹配的恒星大气物理参数自动测量的研究

发布时间：2018-05-03 01:06

本文选题：郭守敬望远镜(LAMOST) + 天体光谱　；参考：《山东大学》2012年硕士论文

【摘要】：人类关于恒星本质的绝大多数知识,几乎都是通过对恒星光谱的研究而得到的。恒星大气物理参数,包括恒星的有效温度、表面重力、化学丰度,是导致恒星光谱差异的重要因素。目前国际上有多种通用的恒星大气物理参数提取算法,利用中低分辨率光谱以及测光数据,在一个相对较窄的参数空间中,提取出相对准确的物理参数。本文主要研究了基于模板匹配的恒星大气参量的自动测量方法,采用的模板库包括理论模板库、实测模板库两大类,将模板匹配算法包括K-最邻近算法、卡方最小化算法、交义相关算法应用到恒星大气物理参数的自动测量中,通过对不同的实测数据的实验表明了这几种方法的有效性。另外还通过实验说明了不同归一化方法以及光谱的信噪比对测量结果的影响。为降低模板匹配的复杂度,本文提出了一种利用人工神经网络(ANN)进行粗估温度缩小匹配模板数的方法,此外还可以将程序部署到并行计算环境中,以进一步提高效率最终在Linux环境下实现程序。本研究的工作介绍本文的主要工作是基于模板匹配的恒星大气物理参数自动测量的研究。LAMOST已经进入先导巡天阶段,即将开始正式巡天,会产生大量光谱,本文的目的是对一维恒星光谱进行处理,利用模板匹配的相关算法,自动获得恒星大气物理参数。本文的工作包括以下几点： 1、提出了一种利用人工神经网络(ANN)进行粗估温度缩小匹配模板数的方法,从而降低模板匹配的复杂度,提高了模板匹配的效率,大大缩短匹配时间。 2、重点研究通过模板匹配方法测量恒星大气物理参数的算法,并通过对不同的实测数据的实验表明了几种模板匹配算法的有效性。 3、通过实验说明了不同归化方法以及光谱的信噪比对测量结果的影响。 4、将程序部署到并行计算环境中，，以进一步提高效率。 5、在Linux环境下用Python语言结合SciPy、NumPy、PyFITS及Matplotlib工具包实现基于模板匹配的恒星大气物理参数自动测量程序
[Abstract]:The vast majority of human knowledge about the nature of stars is almost obtained through the study of stellar spectra. Stellar atmospheric physical parameters, including star effective temperature, surface gravity and chemical abundance, are important factors leading to star spectral differences. At present, there are many universal algorithms for extracting atmospheric physical parameters of stars in the world. Using low and medium resolution spectra and photometry data, relatively accurate physical parameters are extracted in a relatively narrow parameter space. This paper mainly studies the automatic measurement method of stellar atmospheric parameters based on template matching. The template library includes theoretical template library and measured template library. Template matching algorithms include K- nearest neighbor algorithm and chi-square minimization algorithm. The cross-sense correlation algorithm is applied to the automatic measurement of the physical parameters of stellar atmosphere. Experiments on different measured data show the effectiveness of these methods. In addition, the effects of different normalization methods and spectral signal-to-noise ratio on the measurement results are illustrated by experiments. In order to reduce the complexity of template matching, this paper proposes a method of reducing the number of matching templates by using artificial neural network (Ann) to estimate the temperature roughly. In addition, the program can be deployed to parallel computing environment. To further improve the efficiency of the final implementation of the program in the Linux environment. Introduction to the work of this study The main work of this paper is to study the automatic measurement of atmospheric physical parameters of stars based on template matching. LAMOST has entered the stage of leading sky survey, which will produce a large number of spectra soon. The purpose of this paper is to process the spectrum of one-dimensional stars. Using the correlation algorithm of template matching, the atmospheric parameters of stars can be obtained automatically. The work of this paper includes the following points: 1. An artificial neural network (Ann) method is proposed to reduce the number of matching templates, which can reduce the complexity of template matching, improve the efficiency of template matching and greatly shorten the matching time. 2. The algorithm of measuring the physical parameters of stellar atmosphere by template matching method is studied emphatically, and the validity of several template matching algorithms is proved by experiments on different measured data. 3. The effects of different domestication methods and spectral signal-to-noise ratio on the measurement results are illustrated by experiments. In order to further improve efficiency, the program is deployed to parallel computing environment. 5. The automatic measurement program of atmospheric physical parameters of stars based on template matching is realized by using Python language, SciPyNum PyFITS and Matplotlib toolkit in Linux environment.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：P144

【引证文献】