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
Requirements¶
sqlite3
available as a command line programCreate Ocean Networks Canada account and get your API token here
Contents: