Identifying fish sounds of British Columbia with an autonomous audio and video array

The Journal of the Acoustical Society of America(2019)

引用 1|浏览5
暂无评分
摘要
Among the ∼400 marine fish species of British Columbia, only 22 have been reported to be soniferous. However, it is likely due to the lack of examination, as more species are suspected to produce sound. Here we describe how an autonomous audio and video array can identify fish sounds in situ. The array is composed of a collapsible PVC frame, an acoustic recorder, six hydrophones, and two custom-made wide-angle autonomous video cameras mounted on the top and side of the array that collect data continuously for up to 10 days. Fish sounds are automatically detected in the acoustic recordings, the time difference of arrivals between pairs of hydrophones is measured by cross correlation, and the 3-D sound-source location and its uncertainty are estimated using linearized inversion. Simulated annealing optimization was used to define the hydrophone configuration that provides the smallest localization uncertainties. The video recordings are used to assign the species of sound-producing fish localized within the array. The array was deployed at several locations around Vancouver Island and used to define the species, sound characteristics, and source levels of several fish sounds. This new information will help making passive acoustics a viable way to monitor fish in the wild.
更多
查看译文
关键词
fish sounds,autonomous audio,british columbia,video
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要