To improve the geolocation performance of an Unexploded Ordnance (UXO) survey platform, a geodetic Global Positioning System (GPS) receiver was combined with two tactical-grade Inertial Measurement Units (IMUs) and mounted on two types of detection systems. Analysis of data collected for typical trajectories focused on the dual-IMU/GPS pre/post processing using optimal nonlinear estimation together with a Wave Correlation Filter (WCF) and end-matching. Each trajectory of the platforms was estimated by the Extended Kalman Filter (EKF) and the Unscented Kalman Filter (UKF). The WCF was then applied to the two solutions of the platform trajectories derived from each IMU in order to extract the common components in the frequency domain, assuming that uncorrelated components are errors. The remaining bias and trends of the estimated position results were further removed by end-matching IMU solutions and GPS update points. The results of these methods were applied to our field test data to show that the WCF and end-matching can improve position accuracy from 4% to 14% with respect to the Unscented Kalman Smoother (UKS) solution alone.