A Wavelet Transform-based Feature Extraction Pipeline for Elephant Rumble Detection

Abstract

Elephants generate infrasonic vocalisations that traverse through the air for long distances. Utilising this phenomenon, a previous work proposed a system, called Eloc, tolocalise and track elephants in the wild. The Eloc system hasbeen demonstrated to be accurate in calculating the location of infrasonic sources. However, it still lacks the capability toaccurately distinguish elephant infrasonic calls from variousother infrasonic sources using limited computing power on board.Addressing this problem, the work presented in this paperintroduces an approach to distinguish elephant infrasonic callswith a high accuracy on low-resourced hardware. Firstly, asequence of operations are performed to reduce the effect ofnoise in the infrasonic signal captured by an Eloc node. Secondly,a wavelet-based signal reconstruction technique is applied toextract spectral features from the infrasonic signal. Finally, theextracted features are fed to a pre-trained machine learningclassifier to distinguish the infrasonic vocalisations of elephants.The experimental evaluation using Asian elephant (Elephas Maximus Maximus) infrasonic vocalisation datasets demonstrates thatthe proposed approach is capable of accurately distinguishingelephant infrasonic calls on low-resourced hardware platform ofthe Eloc system, with accuracy levels over 82% under varyingenvironmental conditions.