Automatic Visual Analysis of Acoustic Backscatter Time Series for Monitoring Underwater Environments

点击翻译按钮获取中文摘要

The goal of this proposal is to develop new software tools for the automated processing and analysis of underwater acoustic data acquired with echosounders. These new tools will be based upon computer vision and machine learning methods. The potential outcomes include the automatic detection of marine life, such as herring, salmon, and eulachon, as well as of various underwater events and physical phenomena, such as ocean turbulence, presence of suspended sediments, and presence of oil in water. The potential impacts are significant with respect to improving the efficiency and accuracy of species abundance tracking and environmental monitoring. They will enable a paradigm shift from the traditional time-consuming, error-prone manual or semi-automatic acoustic survey data analysis towards novel automatic analysis methods, and thus will reduce processing times, required man-power, and inconsistencies in results.