Using single-camera geometry to perform gyro-free navigation and attitude determination

Abstract

This paper focuses on using an image-based optical sensor (e.g. a digital camera) to estimate relative position and attitude without the explicit dependence on Inertial measurement Unit (IMU) gyro measurements, thus, avoiding sensitivities to gyro drift errors. The method uses a priori known point features and is an alternative to existing resection methods used in photogrammetry that can be easily integrated with previously developed tight optical integration (TOI) method. The key principle involved in this approach is the application of geometric constraints derived from multiple point features and multiple views. The envisioned application of this method is cooperative urban navigation, for ground and airborne vehicles.