Datasets of Odontocete Sounds Annotated for Developing Automatic Detection Methods

• Main Author: Mellinger, David K.
• Other Authors: COOPERATIVE INST FOR MARINE RESOURCES SUTUDIES (CIMRS) NEWPORT OR
• Format: Text
• Language: English
• Published: 2010
• Subjects: Biology; *AQUATIC ANIMALS; ALGORITHMS; DETECTION; SOUND; ENVIRONMENTAL LAW; RECORDING SYSTEMS; WHALES; COMPUTER PROGRAMS; ACOUSTICS; ODONTOCETES; baleen whales
• Online Access: http://www.dtic.mil/docs/citations/ADA535841; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA535841

Environmental laws now require marine research and operations to be conducted so as to minimize impacts on marine mammals and mitigate any adverse impacts caused by that research/operation. Acoustic techniques have become the primary tool used for this task, and have produced vast quantities of data. Automatic call detection software has rapidly become the method of choice for analyzing these data. Most methods for detecting and classifying animal sounds in marine science have been developed for low frequency sounds of baleen whales. However, it is odontocetes (especially beaked whales) that have been most affected by marine operations, particularly of the U. S. Navy, and are of greatest public concern. Therefore, development of high-performance automatic detection methods for odontocetes is now imperative. This report chronicles the progress of this project to collect recordings of beaked whales and other marine mammal species; to annotate these sound files to make them useful to researchers working on automatic call detection and classification; to make the sound files publicly available in an archive on the Internet ("MobySound"); and to develop and test automatic detection algorithms and software, particularly software for autonomous ocean sensing systems such as gliders. The original document contains color images.