Tracking of Multiple Light Sources Using Computer Vision for Underwater Docking

An underwater docking platform can be used as a recharging station for autonomous underwater vehicle (AUV) to prolong its operation. Prior researches indicate that optical sensor is used by an AUV to capture docking station images so as to guide the vehicle towards the station. However, docking using vision is difficult to realize due to the problem of determining the position and orientation of the station with respect to the AUV. Consequently, a tracking system using computer vision is proposed in this paper. Obtained results showed that the tracking method was able to recognize the light sources placed on the docking station which is represented by artificial targets with True Positive Rate of more than 0.9. Furthermore, it was found from this study that as the camera approaching speed towards the target increases, successful recognition decreases. Therefore, an AUV have to propel slowly and steadily to have a successful docking operation.