Are you interested in helping out with some of our cool data analysis? If so give me a call on 021 268 1904 or email andrew@climbingjack.com.
Looking for 4 volunteers – 1 SD card for each person. Around 10 hours of scrolling through sound files. As long as you have a computer that can take an SD card I can help you on line through the rest and find a way to get an SD card to you.
Some file examples are enclosed that come from the recorders that Stu Muir and his son Sandy, shifted around during the spring on the islands out from Aka Aka. Lots of interesting action. Great to see the amount of bittern booming activity, (male mating call).
The bittern are very rare with only around 1,000 in NZ. We are looking to establish some trend monitoring data over time. Stu Muir is working with Fish & Game with much success with a large scale pest control programme in this Nationally Significant Natural Area. It is hoped that as well as bittern, other rare water birds such as Spotless Crake, Fernbird, Banded Rail and our Long Tail Bats will make a big comeback.
clearly a worthy project. But – there must be a better way? Cacophony project? machine learning? If we spend 10hours on the recording, what else is there in there that we’ll miss.
ps – seismic anomaly.
ok – answered my own question.
https://towardsdatascience.com/sound-based-bird-classification-965d0ecacb2b
87% correct species identification from recordings with birdsound. Pretty excellent.
Quicker to listen to the recordings. 🙁
Yeah it’s coming. Tim Hunt at Cacophony has been working in this space but only to detect morepork at the moment. It’s all open source though and I’m certain he’d help anyone to train the AI for bittern.
https://cacophony.org.nz/how-find-morepork