A robust visual servo scheme for underwater pipeline following

Remotely operated vehicles (ROV) are unmanned underwater submersibles for undersea inspection and exploration. This paper presents a vision-based approach for the ROV's underwater pipeline following tasks. The proposed technique overcomes poor image quality of the underwater circumstance including limited brightness, suspended matter in the water, and incomplete pipeline contour due to soils or seaweeds. Edge detection is first applied to locate position of the pipeline with respect to the underwater vehicle. Then, the optical flow method is employed to guide the vehicle tracking the pipeline. Simulation results demonstrate promising performance for underwater pipeline following missions without extra position sensing devices.