Classification of Low-Resolution Stellar Spectra via Template Matching - A Simulation Study

The availability of wavelength- and intensity-calibrated low- resolution spectra directly from spectral survey plates and frames opens the possibility for quantitative spectral classification using modern pattern recognition methods. We investigate the classification power of a template matching technique, which cross-correlates each object spectrum with templates from an atlas of reference spectra. Numerical experiments show that this technique has a high noise-cheating capability and bears great potential for intragalactic as well as extragalactic spectral survey work.