This paper proposes a tightly-integrated Vision/GNSS navigation system for aircraft final approach. It consists of: i) a Kalman filter-based state estimator which handles time-delayed vision measurements by using image-trigger signals of camera device, and ii) an integrity monitoring (IM) function for sensor fault detection and Protection Level calculation. The integrity monitoring function is founded on a batch-realization of Kalman filter. The paper introduces an IM reset mechanism which re-initializes a fault detector regularly without relying on the current state estimation result, in order to remove an influence of past undetected faults on the newly reset detector. The proposed navigation system is tested on real sensor measurements, acquired onboard an unmanned aircraft in flight, with simulated GNSS faults. The test results show an improvement in fault detectability as well as in navigation performance by adding onboard vision information to classical aircraft navigation system.