A Hybrid Marker-Based Camera Tracking Approach in Augmented Reality

The key technological challenge to build effective augmented reality (AR) system is to maintain the accurate and robust tracking between real and computer-generated scene. Vision-based tracking technologies can use passive markers, but they are more computationally demanding and often exhibit erroneous behaviour due to occlusion or numerical instability. Active (sensor-emitter) tracking technologies require powered-device installation, limiting their use to prepared areas. Inertial sensors are completely passive without external devices or targets. However, the drift rates in portable strapdown configurations are too great to practical use. In this paper, we propose a novel camera tracking approach for AR that integrates inertial sensor in marker-based technologies. The complementary nature of these two technologies could compensate for the weakness in each component (marker partly missed). Some experimental results demonstrate the framework effectiveness.