In this work I discuss the necessary steps for deriving photometric redshifts for luminous red galaxies (LRGs) and galaxy clusters through simple empirical methods. The data used is from the Sloan Digital Sky Survey (SDSS). I show that with three bands only ({it gri}) it is possible to achieve results as accurate as the ones obtained by other techniques, generally based on more filters. In particular, the use of the $(g-i)$ color helps improving the final redshifts (especially for clusters), as this color monotonically increases up to $z sim 0.8$. For the LRGs I generate a catalog of $sim 1.5$ million objects at $z < 0.70$. The accuracy of this catalog is $sigma = 0.027$ for $z le 0.55$ and $sigma = 0.049$ for $0.55 < z le 0.70$. The photometric redshift technique employed for clusters is independent of a cluster selection algorithm. Thus, it can be applied to systems selected by any method or wavelength, as long as the proper optical photometry is available. When comparing the redshift listed in literature to the photometric estimate, the accuracy achieved for clusters is $sigma = 0.024$ for $z le 0.30$ and $sigma = 0.037$ for $030 < z le 0.55$. However, when considering the spectroscopic redshift as the mean value of SDSS galaxies on each cluster region, the accuracy is at the same level as found by other authors: $sigma = 0.011$ for $z le 0.30$ and $sigma = 0.016$ for $030 < z le 0.55$. The photometric redshift relation derived here is applied to thousands of cluster candidates selected elsewhere. I have also used galaxy photometric redshifts available in SDSS to identify groups in redshift space and then compare the redshift peak of the nearest group to each cluster redshift (ABRIDGED).
Download