LAMOST二维数据处理方法研究

发布时间：2018-03-10 04:17

本文选题：LAMOST　切入点：光纤光谱　出处：《中国科学技术大学》2010年博士论文　论文类型：学位论文

【摘要】： LAMOST是中科院国家天文台主导研制的具有世界领先水平的大型天文望远镜,与其他同类望远镜相比,它兼具大口径和大视场两个优点,同时它配备了更多数目的光纤和先进的光谱仪,这些技术上的优势将使LAMOST的观测效率和观测深度远远超出其他同类望远镜。二维光纤光谱数据处理系统是LAMOST望远镜的重要组成部分,它对望远镜观测的原始光谱进行处理,获取有科学价值的一维目标天体光谱,供后续天文学者的研究。LAMOST观测的光谱数据量巨大,并且系统的规模和复杂度大大超出同类的望远镜,因此对二维数据处理系统提出了更高的要求。本论文选取LAMOST二维数据处理系统作为研究课题,简要介绍了项目的背景和整体情况,重点分析了二维数据处理的各个流程,着重分析每个处理流程对后续处理的影响。在此基础上,确定了光谱抽取和波长定标两个处理流程作为本论文研究的重点内容；同时,结合作者本人参与LAMOST项目期间所作的具体工作,将实测数据的处理作为另一个研究的重点。本论文的主要研究工作和创新之处有以下几点：研究了基于双高斯径向基函数的光谱抽取算法。首先,介绍了已有的多目标光纤光谱抽谱算法,并分析了这些方法应用于LAMOST光谱抽取存在的不足；随后,根据LAMOST光纤光谱的成像模型,提出了基于径向基函数网络的光谱抽取算法。为了验证该算法的可行性,先是选取简单的单高斯函数作为基函数进行实验。为了更好的满足LAMOST光谱抽取的要求,针对其光纤光谱轮廓本身的特点,提出了双高斯基函数模型,并研究了适当的迭代逼近算法。最后,利用LAMOST和SDSS的实测数据,以及针对性的模拟数据,对算法进行了验证。实验结果表明,所研究的新算法,在抽谱信噪比、抽谱精度、对交叉污染的抑制以及对轮廓非对称性修正等方面都可以满足LAMOST光谱抽取的要求。研究了无需人工干预和先验知识的自动波长定标算法。首先,介绍了多目标光纤光谱波长定标的任务要求和意义；并结合LAMOST光纤光谱本身的特点,分析了现有方法用于LAMOST波长定标的不足之处；然后,根据LAMOST波长定标灯谱局部线性的特点,设计了自动的波长定标算法,详细介绍了该算法的处理步骤。最后,利用LAMOST的实测数据和模拟数据,对自动波长定标算法进行了验证。实验结果表明：自动波长定标算法的精度满足LAMOST二维数据处理的要求；算法的运行时间远远小于人工干预和基于先验知识的波长定标方法；算法能很好地抑制定标灯谱中噪声的影响,具有良好的稳健性。利用实测数据对LAMOST二维光纤光谱数据处理软件的性能进行了检验,并针对实际处理中出现的问题,对二维数据处理软件进行了必要的改进和完善。实验选取了光谱仪调试期间、LAMOST小系统联调期间和LAMOST试观测期间这三个关键节点的实测数据。对光谱仪调试期间实测数据的处理,测试了其主要性能指标,同时根据数据处理结果为光谱仪调试和验收工作提供了积极的建议。对LAMOST小系统联调期间实测数据的处理,检验了二维数据处理的大部分流程,并针对小系统自身的特点,对二维数据处理软件进行了必要的改进和完善；同时也为各个子系统的联调工作提供了有益的参考。对LAMOST试观测期间实测数据的处理,检验了二维数据处理软件的所有流程,并根据数据处理中出现的问题,对二维数据处理软件进行了全面的改进和完善；同时也获得了一批有科学价值的目标天体的红蓝端全波段完整光谱,有利于天文学家通过LAMOST系统开展天文课题的研究。
[Abstract]:LAMOST is a leading national astronomical observatory of Chinese Academy of Sciences has developed the world's leading large astronomical telescope, compared with other similar telescope, it has two large diameter and large field strengths, at the same time it is equipped with a greater number of fibers and advanced spectrometer, these technical advantages make LAMOST observation efficiency and observation depth far beyond other the same telescope.
Spectral data processing system is an important part of the LAMOST telescope, its original spectra of the telescope were processed to obtain one-dimensional object spectra with scientific value of the spectral data of.LAMOST measurements on the amount for subsequent astronomers is huge, and the system size and complexity are far beyond the similar telescope, therefore higher requirements for data processing system.
This paper selects the LAMOST data processing system as a research topic, briefly introduces the background and general situation of the project, analyzed the various processes of two-dimensional data processing, analyzing the influence of each process on the subsequent treatment. On this basis, the spectra extraction and wavelength calibration of two processes as the key content of this paper Study on the practical work done; at the same time, the author participated in the LAMOST project during the processing of measured data as another important research. The main research work and innovation are as follows:
Study on the double Gauss spectra extraction algorithm based on radial basis function. Firstly, introduces the existing multi object Fiber Spectroscopic extraction spectrum methods, and analyzes the shortcomings of these methods are applied to LAMOST spectra extraction there; then, according to the imaging model of LAMOST Fiber Spectroscopic, the spectra extraction algorithm based on RBF network in order to. Verify the feasibility of the algorithm, the first is to select single simple Gauss function as the basis function of experiment. In order to meet the requirements of LAMOST spectra extraction better, according to the characteristics of the spectral profile of the proposed double Gauss function model, and an appropriate iterative approximation algorithm. Finally, using the measured data of LAMOST and SDSS according to the simulation data, and the algorithm was validated. The experimental results show that the new algorithm, the spectral signal-to-noise ratio of pumping, pumping spectral accuracy of cross contamination The requirements of LAMOST spectral extraction can be met in the aspects of suppression and asymmetric profile correction.
Study on the non automatic wavelength without manual intervention and prior knowledge of the calibration algorithm. Firstly, introduces the tasks and meaning of multi object Fiber Spectroscopic wavelength calibration; combined with the characteristics of LAMOST optical fiber itself, analyzes the existing shortcomings of LAMOST method for wavelength calibration; then, according to the characteristics of LAMOST wavelength light spectrum local linear, design the automatic wavelength calibration algorithm, introduces the processing steps of the algorithm. Finally, using the measured LAMOST data and simulated data, the automatic wavelength calibration algorithm was verified. The experimental results show that the automatic wavelength calibration algorithm of LAMOST precision can meet the requirements of two-dimensional data processing; running time the algorithm is far less than the artificial intervention and prior knowledge of the wavelength calibration method based on influence; algorithm can effectively inhibit the noise in the light spectrum, and has good robustness.
The LAMOST spectra of the data processing software performance is tested by using the measured data, and according to the actual process problem of two-dimensional data processing software was improved and perfected. The experiment selected spectrometer during commissioning, LAMOST system debugging and LAMOST data during the three key node test observation during the debugging of the spectrometer. The measured data, the main performance index of the test, at the same time, according to the results of data processing for the commissioning and acceptance of the spectrometer provides positive suggestions. For LAMOST small system alignment during processing of the measured data, the test processes most of two-dimensional data, and according to the characteristics of small system itself. The two-dimensional data processing software was improved and perfected; at the same time also provides a useful reference for the adjustment of other subsystems of LAMOS. The measured data observed during T test, tested all the process of two-dimensional data processing software, and according to the problems in the data processing of the two-dimensional data processing software to conduct a comprehensive improvement; at the same time also received a number of valuable objects of the red and blue side band full spectrum, is conducive to astronomers through the LAMOST system to carry out astronomical research topic.

【学位授予单位】：中国科学技术大学
【学位级别】：博士
【学位授予年份】：2010
【分类号】：TH751

【参考文献】