Multibeam echosounder noise classification using machine learning
Icebergs are large masses of ice drifting in ocean under influence of winds in currents. Icebergs are important indicators of climate, they provide nutrients to the ocean ecosystem, but they are also a threat to shipping and offshore industry. Collecting iceberg shapes helps to model climate, predict iceberg drift, and protect offshore facilities and flowlines laying on the ocean floor. Although, icebergs are fascinating to observe, their larger underwater parts are hidden. It is possible to retrieve the iceberg keel geometry by using an echo-sounder, however, data comes noisy. It takes significant amount of time to filter noise manually. This project will try to apply machine learning algorithms to filter data effectively and detect unnecessary noise in the data. Once the data is clean, iceberg shapes will be determined with higher accuracy, providing researchers and engineers with valuable inputs.
Voir la description complète du projetRocky Taylor;Renat Yulmetov
C-CORE
Engineering
Mining; Professional, scientific and technical services
Memorial University of Newfoundland
Accelerate
