Tehom - Machine Learning on Underwater Acoustics

This package facilitates the creation of machine learning training data on underwater acoustics. While the raw data is available from Marine Cadastre and Ocean Networks Canada (via the onc package), tehom tracks downloads that it conducts and exposes more useful queries for data exploration and sampling. e.g.:

“What hydrophones outside Vancouver have data during April-June, 2017?”

“Where have I downloaded time/geo overlapping data?”

“Give me acoustic snippets as numpy arrays, labeled with whether a container ship was close to the hydrophone that recorded the snippet.”

See example.ipynb in the GitHub repo for an intro to the basic commands



Indices and tables