Login / Signup

A Comparison between Video and Still Imagery as a Methodology to Determine Southern Hairy-Nosed Wombat (Lasiorhinus latifrons) Burrow Occupancy Rates.

Michael J SwinbourneDavid A TaggartBertram Ostendorf
Published in: Animals : an open access journal from MDPI (2018)
Broad-scale abundance estimates of the southern hairy-nosed wombat population use a proxy measure based on counting the number of active burrows, which is multiplied by an index of 'wombats/active burrow'. However, the extant indices were calculated in the 1980s, prior to the use of calicivirus to control rabbits, and used invasive monitoring methods which may have affected the results. We hypothesise that the use of video might provide a logistically simple, non-invasive means of calculating updated indices. To this end, motion-activated, infra-red still and video cameras were placed at various distances outside active wombat burrows in the South Australian Murraylands and Eyre Peninsula regions. The captured imagery was inspected to determine how often the burrow was occupied by one or more wombats, and how effective the cameras were at detecting wombat activity. Video data was clearly superior to the still imagery, with more than twice as many burrow occupancies being positively identified (still: 43%). The indices of wombats/active burrow calculated based on video imagery were: Murraylands: 0.43, Eyre Peninsula: 0.42. 1948 false positive videos were recorded, of which 1674 (86%) occurred between noon and sunset.
Keyphrases
  • high resolution
  • big data
  • mass spectrometry
  • artificial intelligence
  • machine learning
  • deep learning
  • data analysis