This paper aims to demonstrate the capabilities of a minimum hardware configuration for autonomous onboard navigation of LEO satellites consisting of one GPS/GLONASS receiver, one 3-axes magnetometer and landmark data provided by a payload earth observation camera. The paper discusses the overall system architecture and the principles of magnetometer and landmark navigation. A data fusion algorithm based on an Extended Kalman Filter is presented as well as a new algorithm for landmark recognition and matching. Performance results, showing the beneficial capabilities of data fusion, are given for backup magnetometer/landmark operation based on simulation studies for the TUD-Satellite road traffic monitoring demonstration mission Traffic Eye as application reference.