An attitude determination system consists of monocular computer vision, triaxial fiber optic gyroscope and two-axis inclinometer is presented. It aims to meet the demand of the five degrees of freedom air-bearing testbed according to actual situation. Ground computer captures images and recognizes pixel coordinates of indicating lamps on the air-bearing testbed, and then the yaw angle can be calculated by formula which derived in this paper. Tracking differentiator is used to reduce calculation region and can improve processing frequency from 7Hz to 12Hz. Based on Multiplicative Extended Kalman Filtering technique, an attitude determination algorithm is designed by using quaternion. The experiment result proves that attitude of the air-bearing testbed can be esti-mated and corrected accurately by this system.
