Acoustic detection of manatee vocalizations
Christopher Niezrecki, Richard E. Phillips, Michael G. Meyer, Diedrich O. Beusse
Abstract
Christopher Niezrecki, Richard E. Phillips, Michael G. Meyer, Diedrich O. Beusse
Abstract
The West Indian manatee (trichechus manatus latirostris) has become endangered partly because of a growing number of collisions with boats. A system to warn boaters of the presence of manatees, that can signal to boaters that manatees are present in the immediate vicinity, could potentially reduce these boat collisions. In order to identify the presence of manatees, acoustic methods are employed. Within this paper, three different detection algorithms are used to detect the calls of the West Indian manatee. The detection systems are tested in the laboratory using simulated manatee vocalizations from an audio compact disk. The detection method that provides the best overall performance is able to correctly identify approximately 96% of the manatee vocalizations. However, the system also results in a false alarm rate of approximately 16%. The results of this work may ultimately lead to the development of a manatee warning system that can warn boaters of the presence of manatees.
OpenAlex reports 28 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
The West Indian manatee (trichechus manatus latirostris) has become endangered partly because of a growing number of collisions with boats. A system to warn boaters of the presence of manatees, that can signal to boaters that manatees are present in the immediate vicinity, could potentially reduce these boat collisions. In order to identify the presence of manatees, acoustic methods are employed. Within this paper, three different detection algorithms are used to detect the calls of the West Indian manatee. The detection systems are tested in the laboratory using simulated manatee vocalizations from an audio compact disk. The detection method that provides the best overall performance is able to correctly identify approximately 96% of the manatee vocalizations. However, the system also results in a false alarm rate of approximately 16%. The results of this work may ultimately lead to the development of a manatee warning system that can warn boaters of the presence of manatees.
Key concepts: Manatee, Endangered species, Environmental science, Fishery, Ecology, Biology, Habitat