Multi-color classification in the calar alto deep imaging survey

We use a multi-color classification method introduced by Wolf et al. ([CITE]) to reliably identify stars, galaxies and quasars in the up to 16-dimensional color space provided by the filter set of the Calar Alto Deep Imaging Survey (CADIS). The samples of stars, galaxies and quasars obtained this way have been used for dedicated studies which are published in separate papers. The classification is good enough to detect quasars rather completely and efficiently without confirmative spectroscopy. The multi-color redshifts are accurate enough for most statistical applications, e.g. evolutionary studies of the galaxy luminosity function. Also, the separation between stars and galaxies reaches deeper than with morphological criteria, so that studies of the stellar population can be extended to fainter levels. We characterize the dataset presently available on the CADIS 1 h-, 9 h- and 16 h-fields. Using Monte-Carlo simulations we model the classification performance expected for CADIS. We present a summary of the classification results on the CADIS database and discuss unclassified objects. More than 99% of the whole catalog sample at R < 22 (more than 95% at R < 23) are successfully classified matching the expectations derived from the simulations. A small number of peculiar objects challenging the classification is discussed in detail. Spectroscopic observations are used to check the reliability of the multi-color classification (6 mistakes among 151 objects with R < 24). From these, we also determine the accuracy of the multi-color redshifts which are rather good for galaxies (σ_z ≈ 0.03) and useful for quasars. We find that the classification performance derived from the simulations compares well with results from the real survey. Finally, we locate areas for potential improvement of the classification.