Automated passive acoustic detection, classification, localization, and tracking methods applied to recorded data from the Pacific Missile Range Facility, Kauai, Hawaii

Abstract

Automated methods will be reviewed for performing passive acoustic detection, classification, localization, and tracking of some marine mammal species and man-made sources. The methods have been applied to recorded hydrophone data from a large aperture seafloor array at the Pacific Missile Range Facility (PMRF) at Kauai, Hawaii, with some of the methods currently implemented in a real-time system at PMRF. The process consists of custom software both in C + + and Matlab in a 3 or 4-step process. Automated detections of various sounds in specific frequency bands are first performed. In some cases, a classification stage is also performed. The third stage involves model-based localization of the detections or classifications. The fourth stage converts the localizations into individual source tracks. Source tracks are currently generated for fin, sei, Bryde’s, humpback, and sperm whales and mid-frequency active sonar (MFAS) transmissions. Performing this process on marine mammals allows information regarding the movement patterns of the whales while calling, as well as information on the species’ calls (e.g., call rate, frequencies, durations, estimated source levels). By performing similar processes on man-made sources, it is possible to determine some marine mammal responses from proximity of Navy vessels and mid-frequency active sonar sources.