Subsea Structural Monitoring With Machine Vision
Structural monitoring of subsea structures is widely used subsea in the oil and gas industry. Monitoring may be performed to keep track on position, movements, loads and fatigue accumulation. BOPs may have motions sensors installed to monitor the lower flex joint angle and the movements of the BOP itself. Installing subsea equipment like templates or conductors may also require monitoring to ensure the installation is within the set requirements. Motion sensors containing accelerometers and gyroscopes are commonly used for these applications. These sensors are well suited for accurately determining inclination, but the sensors estimate displacements poorly. An alternative approach for such measurements is to use machine vision. The machine vision tracks selected features from a video feed and calculates position, displacement and inclination directly. A benefit of this monitoring technique is that it does not require any hardware on the actual structure, only a camera and a tracking software. Two video tracking algorithms are proposed for tracking. One marker based algorithm that require a pre-installed visual marker, and a markerless algorithm that uses natural features on the structure. The article presents the accuracy and project examples of in-field use of these algorithms.