Very metal-poor (VMP, [Fe/H]<-2.0) stars serve as invaluable repositories of insights into the nature and evolution of the first-generation stars formed in the early galaxy. The upcoming China Space Station Telescope (CSST) will provide us with a large amount of spectral data that may contain plenty of VMP stars, and thus it is crucial to determine the stellar atmospheric parameters ($T_{\textrm{eff}}$, $\log$ g, and [Fe/H]) for low-resolution spectra similar to the CSST spectra ($R\sim 200$). This study introduces a novel two-dimensional Convolutional Neural Network (CNN) model, comprised of three convolutional layers and two fully connected layers. The model’s proficiency is assessed in estimating stellar parameters, particularly metallicity, from low-resolution spectra ($R \sim 200$), with a specific focus on enhancing the search for VMP stars within the CSST spectral data. We mainly use 10 008 spectra of VMP stars from LAMOST DR3, and 16 638 spectra of non-VMP stars ([Fe/H]>-2.0) from LAMOST DR8 for the experiments and apply random forest and support vector machine methods to make comparisons. The resolution of all spectra is reduced to $R\sim200$ to match the resolution of the CSST, followed by pre-processing and transformation into two-dimensional spectra for input into the CNN model. The validation and practicality of this model are also tested on the MARCS synthetic spectra. The results show that using the CNN model constructed in this paper, we obtain Mean Absolute Error (MAE) values of 99.40 K for $T_{\textrm{eff}}$, 0.22 dex for $\log$ g, 0.14 dex for [Fe/H], and 0.26 dex for [C/Fe] on the test set. Besides, the CNN model can efficiently identify VMP stars with a precision rate of 94.77%, a recall rate of 93.73%, and an accuracy of 95.70%. This paper powerfully demonstrates the effectiveness of the proposed CNN model in estimating stellar parameters for low-resolution spectra ($R\sim200$) and recognizing VMP stars that are of interest for stellar population and galactic evolution work.

Each CCD of LAMOST accommodates 250 spectra, while about 40 are used to observe sky background during real observations. How to estimate the unknown sky background information hidden in the observed 210 celestial spectra by using the known 40 sky spectra is the problem we solve. In order to model the sky background, usually a pre-observation is performed with all fibers observing sky background. We use the observed 250 skylight spectra as training data, where those observed by the 40 fibers are considered as a base vector set. The Locality-constrained Linear Coding (LLC) technique is utilized to represent the skylight spectra observed by the 210 fibers with the base vector set. We also segment each spectrum into small parts, and establish the local sky background model for each part. Experimental results validate the proposed method, and show the local model is better than the global model.

As an integral component of the LAMOST Experiment for Galactic Understanding and Evolution (LEGUE; Deng et al. 2012), the LAMOST Galactic anti-center spectroscopic survey (Liu et al. in preparation) will survey over three thousand square degree sky area centered on the Galactic anti-center (150d ≤ l ≤ 210d, −30d ≤ b ≤ +30d) and obtain low resolution (R ~ 1800) optical spectra for a statistically complete sample of more than three million stars down to a limiting magnitude of 18.5 in r band, distributed in a spatially contiguous area and probing a significant volume of the Galactic thin/thick disks, halo and their interface. Sample stars of the LAMOST survey of the Galactic anti-center are derived from a recently completed CCD imaging photometric survey utilizing the newly built 1.0/1.2m Schmidt Telescope at the Xuyi Station of the Purple Mountain Observatory. The Xuyi imaging survey (Yuan et al., in preparation; Zhang et al. 2012) provides high quality photometry (~2 per cent) in the SDSS g, r and i bands and astrometry (~0.1 arcsec) for about a hundred million stars down to a limiting magnitude of about 19 (10 sigma) for over six thousand square degree sky area (3h ≤ RA ≤ 9h, −10d ≤ Dec ≤ +60d) that envelopes the LAMOST spectroscopic survey area of the Galactic anti-center, plus an extension to the M 31 and M 33 region.

This Digital Sky Survey of the Galactic Anti-center (DSS-GAC) with the Xuyi Schmidt and LAMOST telescopes will yield for the first time optical photometry and spectra for millions of stars in the Galactic disk(s), the defining component of the Milky Way as a typical spiral galaxy that contains most Galactic baryonic material and angular momentum. DSS-GAC will deliver classification, extinction, radial velocity and stellar parameters (Teff, log g, [Fe/H], probably also [α/Fe], and in some cases, [C/Fe]), for each sample star. Together with the accurate proper motions and distances to be obtained with the forthcoming GAIA mission, DSS-GAC offers a unique opportunity for major breakthroughs in studies of the Galactic structure, formation and evolution. In particular, DSS-GAC will generate a huge data set to 1) study the stellar populations, chemical composition and kinematics of the thin and thick disks and their interface with the halo; 2) Understand how resilient galaxy disks are to gravitational interactions/perturbations and study the temporal and secular evolution of the disks; 3) identify tidal streams and debris of disrupted dwarfs and clusters; 4) probe the gravitational potential and dark matter distribution; 5) map the three-dimensional distribution and extinction of the interstellar medium; 6) search for rare objects (e.g. stars of peculiar chemical composition, hyper-velocity stars); and 7) ultimately advance our understanding of the formation and evolution of stars and galaxies.

Following a two-year commissioning, the LAMOST pilot survey was initiated in October, 2011 and completed in June, 2012. In total, about 370,000 spectra of 270,000 stars have been obtained for DSS-GAC, with 70 per cent of the spectra reaching a spectral S/N ratio per resolution element at 7150 Å higher than 20. The formal LAMOST DSS-GAC survey will commence in September, 2012, and is expected to complete in five years.

This paper describes the 2-dimensional data reduction of the LAMOST pilot survey.