Uncalibrated Visual Servo for the Remotely Operated Vehicle

In this paper, an image-based uncalibrated visual servo scheme is proposed for image tracking tasks in highly disturbed environment, such as a remotely operated vehicle performing observation or investigation influenced by severe underwater current. Under the conditions of unknown target model and camera parameters, the control framework applies the scale invariant feature transform (SIFT) technique to extract image features. Furthermore, a robust adaptive control law is implemented to overcome the effect caused by camera parameters. Then by using three different types of camera's motion: pan, tilt, and zoom to achieve the visual servo mission by maintaining the target always at the central position on the image plane.