Importance

Loading Importance Indicators...

Abstract

Disclosed embodiments provide techniques for automated passive acoustic monitoring with machine learning. An acoustic sensor is accessed. The acoustic sensor includes an embedded acoustic controller which hosts a machine learning model. The acoustic sensor is coupled to one or more hydrophones. The acoustic sensor is deployed in a body of water and is submerged. The acoustic sensor can enter a sleep mode. The hydrophones receive an underwater audio signal. The audio signal can be associated with an acoustic pressure. The acoustic sensor can be woken from sleep when the acoustic pressure is above a pressure threshold. The machine learning model classifies a predicted source of the underwater audio signal. The classifying can be based on filtering the underwater audio signal for a first frequency band associated with a source of interest. The predicted source is reported to a user using a communications device.

Loading the Abstract Image...

First Claim