Obstacle detection and localization relative to surrounding objects or structures are important for safe vehicle navigation. This capability is particularly useful for unmanned surface vessels (USVs) operating near large structures that will block GPS signals. This study proposes a relative navigation approach, which allows estimating the vehicles position under bridges where GPS signals are not available by building a parameterized map of bridge pier structures in the framework of simultaneous localization and mapping (SLAM). The feasibility of the algorithm is demonstrated in outdoor experiments.