Camera-assisted AUV navigation in fish cages
MetadataVis full innførsel
The usage of Remotely Operated Vehicles (ROVs) for inspection of marine structures such as fish nets is today an essential part of the growing offshore industry. In fish farms, the ROV operator has to spend most of the time relatively close to the net, witch can be demanding in rough seas. The ability for an semi autonomous ROV to effectively navigate within the cage and guarantee a complete inspection requires new tools, and dynamic positioning (DP) improvements. This thesis will use the power of today s camera and gpu technology to solve this problem. Two different approaches have been tried, firstly a SLAM algorithm with a EKF for assisted odometry support was tried. Due to hardware problems and a lack of meaningful results from the tests it was abounded for a different method using keypoint matching and image stitching. Time became a limiting factor for the second approach, but the foundation for a working prototype has been made.