Automated acoustic detection of submerged hydrocarbon plumes produced by seeps and spills

Nazareth, Ginelle Claire (2024) Automated acoustic detection of submerged hydrocarbon plumes produced by seeps and spills. Masters thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (18MB)

Abstract

Assessing the impact of an oil spill is especially challenging when it forms a neutrally buoyant submerged plume, as in the case of the Deepwater Horizon well blowout. Autonomous Underwater Vehicles (AUVs) have proven to be an effective platform for detecting, tracking, and sampling these plumes due to their adaptive response capabilities. However, efforts to test AUV spill detection technology at naturally occurring seeps are hindered by challenges posed by seep environments. This thesis consequently focuses on addressing seep-site challenges by enhancing automated acoustic detection of submerged plumes. It proposes a novel plume detection algorithm for forward-looking sonars consisting of five steps: range-gating, segmentation, grid conversion, clustering, and georeferencing. Due to the computational complexity of the clustering step, a custom "block clustering" algorithm for image data is developed to meet real-time processing requirements. A playback test using field trials data collected at Holyrood Bay demonstrates that the plume detection algorithm successfully identifies high-density clusters. Furthermore, the block clustering is consistently faster than the benchmark algorithm and produces clustering results that are visually more intuitive. The plume detection algorithm was implemented on Memorial University's Explorer AUV and utilized during trials to adaptively detect and sample the Scott Inlet seeps in Baffin Bay.

Actions (login required)

View Item