Identification and similarity of individual clicks for abundance estimation of cetaceans
* Presenting author
The assessment of the number of active cetacean individuals in a specified location is essential for the protection and conservation of the marine ecosystem. Current methods for the abundance estimation of cetaceans use passive acoustic monitoring techniques to identify the number of active individuals where visual surveys are not possible. In this work, the echolocation clicks of different whale species are studied to identify statistical characteristics of the calls which can aid in the estimation of the number of active individuals in a recording. Signals are filtered on its occupied bandwidth and harmonic content of the signals is discarded before extracting the click segments. Possible reflections and false detections are identified using a decision tree based on the crest factor and the spectral flatness of the clicks. Several features in the time and frequency domains are then extracted from these click segments so as to find the features which can identify similarities between consecutive clicks. Different click trains are then extracted which help estimate the number of individuals of a particular species in the recordings. The method is evaluated using signals with different species including Pilot Whales, Blainville's Beaked Whales, and Harbour Porpoises.