主催: 一般社団法人 日本機械学会
会議名: ロボティクス・メカトロニクス 講演会2019
開催日: 2019/06/05 - 2019/06/08
The purpose of this study is detection the call of Gallirallus okinawae from recording data for a long time. Gallirallus okinawae are Classified as Critically Endangered which natural monument and Endemic species of Okinawa. Gallirallus okinawae has been living in the forest without flying power, therefore, ongoing investigation of a long period of time is extremely difficult. Therefore, we have investigating of habitat throughout the year from recordings of Gallirallus okinawae sing in the fixed point. In this paper, we proposed a method of detecting a Gallirallus okinawae sing by performing the processing band-pass filter, the Hilbert transform, the envelope detection, the average rate of change calculated, the zero-crossing count for the digital sound data. I including the Hilbert transform into the original algorithm, the rate of accurate detection increased from 46% to 76%.