In this paper we show how to carry out robust place recognition using both near and far information provided by a stereo camera. Visual appearance is known to be very useful in place recognition tasks. In recent years, it has been shown that taking geometric information also into account further improves system robustness. Stereo visual systems provide 3D information and texture of nearby regions, as well as an image of far regions. In order to make use of all this information, our system builds two probabilistic undirected graphs, each considering either near or far information. Inference is carried out in the framework of conditional random fields. We evaluate our algorithm in public indoor and outdoor datasets from the RAWSEEDS project and in an outdoor dataset obtained at the MIT campus. Results show that this combination of information is very useful to solve challenging cases of perceptual aliasing.