The Research On Two Dimensional Data Reduction For LAMOST

LAMOST is a world's leading large astronomical telescope independently developed by National Astronomical Observatories, CAS. Compared with other similar telescopes, it has two advantages as large diameter and large field. Moreover, it is equipped with a greater number of fibers and advanced spectrographs. These technical advantages make the observation efficiency and observation depth of LAMOST far beyond the other similar telescopes.Two-dimensional spectra reduction system is an important part of LAMOST. It processes the raw spectra to obtain one-dimensional object spectra with scientific values, and then hands to astronomers for the follow-up study. Since the spectral data observed by LAMOST is very massive and the system size and complexity are far beyond the similar telescopes, a higher requirement is set to the two-dimensional spectra reduction system.The two-dimensional spectra reduction system of LAMOST is studied in this thesis. A brief description of the project's background is introduced, and relationship of the processing of the two-dimensional spectra reduction is analyzed, especially the impact of foregoing processing to the follow-up ones. On this basis, the spectra extraction and wavelength calibration are selected as the emphasis in this study. At the same time, according to the practical work done during the project, the actual spectra reduction is chosen as another important study. The main work and innovations are listed as follows:A spectra extraction algorithm based on double Gaussian radial function is developed. First, the existing spectra extraction algorithms for multi-object fiber spectra are introduced, and the shortcomings of these algorithms in the application to LAMOST are analyzed. Subsequently, according to the imaging model of LAMOST fiber spectroscopic, the spectra extraction algorithm based on radial basis function networks is proposed. In order to verify the feasibility of the algorithm, a simple single-Gaussian function is chosen as radial function. In order to better meet the requirements of spectra extraction, the double Gaussian function model is proposed according to the spectral profile of LAMOST, and an appropriate iterative approximation is developed. Finally, experiments are implemented on the LAMOST and SDSS data as well as targeted simulated data to evaluate the performance of the new algorithm. The experimental results show that the new algorithm can meet the requirements of LAMOST spectrum extraction in the respects of the signal to noise ratio, the precision, the inhibition of cross-talk and the correction to the non-symmetry profile.An automatic wavelength calibration algorithm without manual interaction or prior knowledge is studied. First, the tasks and meaning of wavelength calibration for multi-objective fiber spectra is introduced, and then the limitations of the existing wavelength calibration methods in the application of LAMOST is analyzed according to characteristics of LAMOST spectra. Subsequently, the automatic wavelength calibration algorithm is developed according to local linear features of the arc spectra, and detailed processing steps of the algorithm are introduced. Finally, experiments are performed on actual data of LAMOST as well as simulated data in order to test the performance of the proposed algorithm. The results of experiments show that the accuracy of the algorithm can meet the requirements of LAMOST, the running time is far less than the methods with manual interaction or based on prior knowledge, and the algorithm has good robustness so that it can well inhibit the impact of noise in the spectra.Using the actual spectra observed by LAMOST, the performance of data reduction software is tested and necessary improvement is made. The actual spectra are chosen from three crucial periods such as the spectrometer debugging, the part LAMOST testing and LAMOST test observation. The experiments on spectra of spectrometer debugging, not only tested the main performance of the spectrometer, but also provide a positive recommendation to the commissioning and acceptance of the spectrometer. The experiments on spectra of part LAMOST, tested most of the processing and made necessary improvement to the data reduction software according to its own characteristics. Meanwhile, a useful reference is provided to the adjustment of other subsystems. The experiments on spectra observed by LAMOST during the test observation made a comprehensive test to the two-dimensional data reduction software. During this period, the two-dimensional data processing software have been fully improved and perfected according to the experimental results. Furthermore, a number of scientific spectra in the entire wavelength range (blue and red band) are acquired, which is helpful for astronomers to carry out astronomical research projects base on LAMOST.