Automated 1D Spectral Processing For LAMOST

As an ancient science, astronomy played an import part since the beginning of human civilization. With the great improvement of observing equipment and data collection capabilities, astronomical observations come into the avalanche era. The optical spectral of celestial bodies contain a great deal of information. From the spectra of galaxies, we can get their distance, composition, distribution and motion etc. From the stellar spectra,their distribution, motion, luminosity, temperature and its chemical composition etc. can be obtained. Furthermore, we can discover some special, rare and unknown objects or phenomena which will lead to human new understanding of celestial bodies.With the launch of LAMOST Survey, tens of thousands of spectra of celestial bodies will be collected in each observation night and totally about 10,000,000 spectra will be obtained. It's difficult to deal such massive spectra by hand, hence some algorithms should be developed to process the spectra automatically.In the LAMOST 1D spectra processing Pipeline, the Stellar Paramater moudle (LAMOST₁DSP) provide physical parameters which are useful in LEGUE.The discovered spectra of special, rare, unknown celestial objects will be large samples for astronomy research.The work consists of two parts:(1) The design and implementation of LAMOST automated 1D Stellar atmospheric Parameters measurement system (LAMOST 1DSP), including the spectral preprocessing and parameter measurement modules which support the expansion of parameter measurement. And now Latter module has integrated SSPP, UlySS packages and the PLS method. There will be more methods integrated in the future work. The system is implemented using Python, GTK and multi-threaded programming. (2) Study some data mining algorithms to discover spectra to find specific objects. The algorithms are divided into two parts:supervised methods and un supervised methods. Research methods include random forest algorithm and genetic algorithm etc.