Detection and Identification of Marine Mammals in Passive Acoustic Recordings from SCORE using a Visual Processing Approach Established for HARP Data
Abstract
A visual processing approach developed for analyzing passive acoustic recordings of marine mammal vocalizations collected on a High-frequency Acoustic Recording Package (HARP) is applied to acoustic data collected through three hydrophones at the Southern California Offshore Range (SCORE) on a Naval Postgraduate School recording system. Temporally overlapping datasets collected in proximity to one another are examined with the expectation that vocalizations from species that normally inhabit this region (resident or transient) were recorded on both systems. The analysis process relies on determination of invariant and distinctive features of marine mammal vocal elements to classify mammal sources. Vocalization features used to identify specific sources in the HARP data appear modified in the SCORE data. We examine how the technical components and recording parameters of the SCORE recording system affect the received acoustic signatures of odontocetes to determine how the visual processing protocols applied to HARP data can be adapted for application to SCORE data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 2012
- Accession Number
- ADA562978
Entities
People
- Elizabeth A. Harris
Organizations
- Naval Postgraduate School