Automatic detection of marine mammals using information entropy

Abstract

This article describes an automatic detector for marine mammal vocalizations. Even though therehas been previous research on optimizing automatic detectors for specific calls or specific species,the detection of any type of call by a diversity of marine mammal species still poses quite achallenge—and one that is faced more frequently as the scope of passive acoustic monitoring studiesand the amount of data collected increase. Information Shannon entropy measures the amount ofinformation in a signal. A detector based on spectral entropy surpassed two commonly useddetectors based on peak-energy detection. Receiver operating characteristic curves were computedfor performance comparison. The entropy detector performed considerably faster than real time. Itcan be used as a first step in an automatic signal analysis yielding potential signals. It should befollowed by automatic classification, recognition, and identification algorithms to group and identifysignals. Examples are shown from underwater recordings in the Western Canadian Arctic. Calls ofa variety of cetacean and pinniped species were detected.