In this work, we propose the reference signal-aided channel estimation approach which is a crucial requirement to address the realistic performance of spatial media-based modulation (SMBM) systems where the digitally controllable radio frequency mirrors are deployed on the multiple transmit antennas. Accordingly, least squares and linear minimum mean square error-based channel estimation schemes are proposed for MBM-based systems for the first time in the literature where former studies mainly assume either perfect channel state information or an error model on channel coefficients. In addition, corresponding symbol detection performance is studied. To measure the efficiency of the proposed channel estimation approaches, the theoretical upper bounds on average bit error rate are derived and shown to be well overlapped with the computer simulations for the medium and high signal-to-noise ratio regions. This study is important due to the implementation of channel estimation as well as the theoretical derivation of detection bounds for SMBM communication systems. (C) 2021 Elsevier B.V. All rights reserved.