Bird song recognition from automatic recorders

Identifying birds from their calls is potentially extremely useful since birds call far more often than they are seen and many bird species are threatened or endangered of extinction. However, while recording birds using autonomous recorders placed where they live can be seen as easy, birds rarely call close to the microphone and there are lots of other sounds that are also recorded (including wind, rain, other birds, other animals, and anthropogenic noise).
The challenge is to reliably recognise the bird calls under these conditions and to infer the number of birds present in an area from the number of calls recorded. This requires a combination of signal processing, machine learning, study of individual bird species and statistical analysis.
We are seeking a PhD student (supported by Te Punaha Matatini – The New Zealand Centre of Research Excellence for Complex Systems, Data and Networks and Marsden Fund, Royal Society of New Zealand). PhD’s research will contribute to the AviaNZ project (Https: http://seat-web4.massey.ac.nz/avianz/) and research by the Behavioural Ecology and Conservation Group (http://www.massey.ac.nz/behaviour-conservation-group).