Observation support system of an ROV for underwater archaeology

In this paper, we propose an observation support system of a Remotely Operated Vehicle (ROV) for underwater archaeology. In general, it is difficult for unskilled users to drive an ROV for underwater observation because a camera mounted on an ROV always moves due to external disturbances and objects easily move out of the field of view (FOV) of the camera. Once an object is out of the FOV, users cannot easily find out it again. The observation support system is used to help unskilled users to drive an ROV. To realize the proposed system, object detection based on image processing is a key component. We apply the SURF (Speeded Up Robust Features) algorithm to detect archaeological objects, and investigate the performance of the algorithm using video images recorded by our ROV in an underwater archaeological site.