基于線指數(shù)的恒星大氣物理參數(shù)提取方法的研究
發(fā)布時(shí)間：2018-08-24 12:10
【摘要】：現(xiàn)在國際上的大口徑兼大視場望遠(yuǎn)鏡有美國的Sloan數(shù)字巡天望遠(yuǎn)鏡,英澳天文臺的2dF巡天望遠(yuǎn)鏡,我國的LAMOST巡天望遠(yuǎn)鏡等。它們將得到海量的光譜數(shù)據(jù)。通過觀測獲得恒星的光譜,不僅能夠確定其大氣參數(shù)和空間分布,還能夠結(jié)合年齡和運(yùn)動學(xué)信息,得到銀河系不同星族的大氣參數(shù),從而為銀河系的形成、結(jié)構(gòu)和演化模型提供準(zhǔn)確的約束條件。將恒星演化模型與觀測結(jié)果進(jìn)行對比,還能有效地追蹤銀河系自形成以來的演化歷史、理解核合成理論,并且對現(xiàn)有的宇宙模型作出檢驗(yàn),促使人們對宇宙演化有更新的認(rèn)識。恒星大氣參數(shù)分析是探索恒星、銀河系、甚至宇宙演化的一種基本途徑。如此海量的光譜數(shù)據(jù)對光譜的快速有效處理提出了更高的要求。 恒星大氣物理參數(shù)(有效溫度、表面重力、化學(xué)豐度)是導(dǎo)致恒星光譜差異的主要因素。恒星大氣物理參數(shù)的自動測量是LAMOST等大規(guī)模巡天望遠(yuǎn)鏡所產(chǎn)生的海量天體光譜數(shù)據(jù)自動處理中一個(gè)重要研究內(nèi)容。 Lick指數(shù)是一個(gè)相對較寬的光譜特性,以每個(gè)線指數(shù)最突出的吸收線命名。這種指數(shù)可以忽略流量的校正和紅移的錯(cuò)誤并對具有更高的信噪比S/N,這讓Lick線指數(shù)成為了測量大氣物理參數(shù)的一種理想方法。 本課題利用Lick線指數(shù),根據(jù)光譜的海量特點(diǎn),分別通過線性回歸、人工神經(jīng)網(wǎng)絡(luò)、支持向量機(jī)的方法來設(shè)計(jì)大氣物理參數(shù)測量的高效算法。該方法首先對Kurucz訓(xùn)練數(shù)據(jù)進(jìn)行篩選,然后利用三種算法對其進(jìn)行訓(xùn)練,找到最佳模型。最后與DR8測試數(shù)據(jù)的相應(yīng)參數(shù)進(jìn)行對比測試。結(jié)果表明,該方法能有效準(zhǔn)確地從低分辨率光譜中測定恒星大氣物理參數(shù)。實(shí)驗(yàn)結(jié)果證明,與目前的方法相比,該方法能夠有效地估計(jì)恒星大氣參數(shù)且速度更快,準(zhǔn)確度更高。 總的來說,利用Lick線指數(shù)來進(jìn)行大氣物理參數(shù)的預(yù)測是具有可行性的。對于本文中的三種算法模型有待于進(jìn)行進(jìn)一步的研究,使其達(dá)到一個(gè)更好的擬合效果,從而可以將該算法更好的應(yīng)用于像LAMOST一樣的巡天項(xiàng)目中去。
[Abstract]:At present, the international large caliber and large field of view telescopes include the Sloan Digital Sky Survey Telescope of the United States, the 2dF Survey Telescope of the British and Australian Astronomical Observatory, the LAMOST Survey Telescope of China and so on. They will get a great deal of spectral data. The spectra of stars can not only determine the atmospheric parameters and spatial distribution, but also combine the age and kinematics information to obtain the atmospheric parameters of different star families in the Milky way, thus contributing to the formation of the Milky way. The structure and evolution model provide accurate constraints. By comparing the stellar evolution model with the observational results, we can effectively trace the evolution history of the Milky way since its formation, understand the nuclear synthesis theory, and test the existing cosmic models, thus promoting a new understanding of the evolution of the universe. Stellar atmospheric parameter analysis is a basic way to explore the evolution of stars, galaxies, and even the universe. Such a large amount of spectral data put forward a higher demand for fast and effective processing of spectrum. Stellar atmospheric physical parameters (effective temperature, surface gravity, chemical abundance) are the main factors leading to star spectral differences. The automatic measurement of stellar atmospheric physical parameters is an important research content in the automatic processing of massive celestial body spectral data produced by LAMOST and other large-scale survey telescopes. The Lick index is a relatively wide spectral characteristic. Named after the absorption line that is most prominent in each line index. This index can ignore the error of flow correction and redshift and has a higher SNR S / N, which makes the Lick line index an ideal method for measuring the physical parameters of the atmosphere. In this paper, Lick line exponent is used to design an efficient algorithm for measuring atmospheric physical parameters by means of linear regression, artificial neural network and support vector machine, according to the magnanimity of spectrum. Firstly, the Kurucz training data are filtered, then three algorithms are used to train it to find the best model. Finally, it is compared with the corresponding parameters of DR8 test data. The results show that this method can effectively and accurately measure the atmospheric physical parameters of stars from low resolution spectra. The experimental results show that the proposed method can effectively estimate the atmospheric parameters of stars faster and more accurately than the present method. In general, it is feasible to use Lick line index to predict atmospheric physical parameters. The three algorithms in this paper need to be further studied in order to achieve a better fitting effect, so that the algorithm can be better applied to the survey projects like LAMOST.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：P144;TP18
本文編號：2200781
[Abstract]:At present, the international large caliber and large field of view telescopes include the Sloan Digital Sky Survey Telescope of the United States, the 2dF Survey Telescope of the British and Australian Astronomical Observatory, the LAMOST Survey Telescope of China and so on. They will get a great deal of spectral data. The spectra of stars can not only determine the atmospheric parameters and spatial distribution, but also combine the age and kinematics information to obtain the atmospheric parameters of different star families in the Milky way, thus contributing to the formation of the Milky way. The structure and evolution model provide accurate constraints. By comparing the stellar evolution model with the observational results, we can effectively trace the evolution history of the Milky way since its formation, understand the nuclear synthesis theory, and test the existing cosmic models, thus promoting a new understanding of the evolution of the universe. Stellar atmospheric parameter analysis is a basic way to explore the evolution of stars, galaxies, and even the universe. Such a large amount of spectral data put forward a higher demand for fast and effective processing of spectrum. Stellar atmospheric physical parameters (effective temperature, surface gravity, chemical abundance) are the main factors leading to star spectral differences. The automatic measurement of stellar atmospheric physical parameters is an important research content in the automatic processing of massive celestial body spectral data produced by LAMOST and other large-scale survey telescopes. The Lick index is a relatively wide spectral characteristic. Named after the absorption line that is most prominent in each line index. This index can ignore the error of flow correction and redshift and has a higher SNR S / N, which makes the Lick line index an ideal method for measuring the physical parameters of the atmosphere. In this paper, Lick line exponent is used to design an efficient algorithm for measuring atmospheric physical parameters by means of linear regression, artificial neural network and support vector machine, according to the magnanimity of spectrum. Firstly, the Kurucz training data are filtered, then three algorithms are used to train it to find the best model. Finally, it is compared with the corresponding parameters of DR8 test data. The results show that this method can effectively and accurately measure the atmospheric physical parameters of stars from low resolution spectra. The experimental results show that the proposed method can effectively estimate the atmospheric parameters of stars faster and more accurately than the present method. In general, it is feasible to use Lick line index to predict atmospheric physical parameters. The three algorithms in this paper need to be further studied in order to achieve a better fitting effect, so that the algorithm can be better applied to the survey projects like LAMOST.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：P144;TP18
【參考文獻(xiàn)】
相關(guān)期刊論文 前3條
1 Philippe Prugniel;;Automatic determination of stellar atmospheric parameters and construction of stellar spectral templates of the Guoshoujing Telescope (LAMOST)[J];Research in Astronomy and Astrophysics;2011年08期
2 ;The Large Sky Area Multi-Object Fiber Spectroscopic Telescope (LAMOST)[J];Research in Astronomy and Astrophysics;2012年09期
3 ;Data release of the LAMOST pilot survey[J];Research in Astronomy and Astrophysics;2012年09期
，本文編號：2200781
本文鏈接：http://sikaile.net/kejilunwen/tianwen/2200781.html
教材專著
