A Real-Time Pipeline Tracking Using a Forward-Looking Sonar

Abstract

Abstract The pipeline inspection is one of the most critical tasks in the oil and gas industries. The operation of a remotely operated vehicle (ROV) for the pipeline inspection is very costly since it requires both a human operator and the support vessel to which an ROV must be tethered. Furthermore, this tether also limits the surveillance ability of an ROV. Thus, the use of autonomous underwater vehicles (AUVs) can greatly reduce the operation cost of pipeline inspection since an AUV can freely and automatically navigate without tethering to support vessel. In order for an AUV to navigate automatically under the limited vision in the underwater environment, the forward-looking sonar (FLS) is the essential part since the quality of images captured by an optical camera degrade significantly by the turbid water and the low-light condition whereas images produced from an acoustic sonar As a result, in this work, we are interested in the use of a 2-D multi-beam Forward-Looking Sonar (FLS) to capture echo images on a frontal vision since acoustic signals can pass through tiny particles in turbid medium water [1]. This deployment ensures that an image of pipelines can be captured regardless of low underwater visibility conditions. Other types of sonars that are regularly used in underwater visual inspection include the side scan sonar (SSS), synthetic aperture sonar (SAS), and multi-beam echo sounder (MBES). There are a limited number of works that focus on the pipeline detection problem. An optical camera was employed in [2] for visual feedback and heading control, SSS and MBES were used in [3], [4] and [5] for pipeline tracking, and, in [6], FLS was used for linear object detection and tracking.