Development of underwater object tracking system with camera as visual feedback

Remotely Operated Vehicle (ROV) has been widely used and become important tool in exploring underwater. ROV is used to take underwater samples, maintenance of deep sea pipes and investigate transportation accident at sea. Object tracking is one of the focus of research that continues to be developed with many methods. Color segmentation algorithm fails to detect object in complicated environment because of many factors. In this paper we used a new strategy to improve the ability of color segmentation algorithm that can estimate the position of the object using a combination of color segmentation and optical flow to mark and detect the shape of the object. The developed method works in real time to transmit position of the object to move the ROV direction. The experimental results of the performance have been tested in the swimming pool during the day with various tests which results will be described in this paper. This research focuses on making the automatic object tracking system in the water by using the camera as a visual sensor. The designed system can detect the position of an object by performing image processing. Tracking the object using a combination of color segmentation and optical flow so that ROV can move autonomously closer to the selected object.