Hello,
These questions continues the theme of my previous questions regarding how to video and analyse insects in agricultural contexts on a budget in somewhere like Malawi. Thanks to everyone who has contributed so far, it has been really useful and I am getting much closer to a pilot design and method that I can trial in a scoping visit this May.
I am considering recording continuous video of baits (such as seeds, seedling or insects) to see what is removing them (could be mammals, birds, insects) using a security camera mounted directly above the bait. I'd then run this video through software to detect motion (this way I can adjust sensitivity as appropriate, which might be crucial for small insects) and identify animals into broad groups (ants, beetles, mice, birds) or better.
My set up will probably be a 12V 90 AH car battery powering a 4-8 channel DVR with bandwidth of 2-8 Mb/S with affordable active IR security cameras. Such cameras typically have 1280*1024 (1.0 MP) or 704*480 (4CIF) resolutions.
My current design considerations/questions are:
1) Assuming a clear image, what resolution (pixels/mm) might be needed to identify a relatively small insect such as a driver ant worker (20-25 mm2) from other insects such as beetles?
2) What is the best frame rate for retaining identification and behaviour data (insects visiting baits or damaging seedlings) whilst minimizing the load on DVR bandwidth?
3) What is the resolution of active IR security cams at night vs. day?
4) I have tried searching but I can't find any useful info on minimum focal distance for security cam lenses. So does anyone know how can I calculate or find out minimum focal distance and depth of field for a typical CCTV bullet cam + various lenses?
Any thoughts and advice on these would be very welcome.
Many thanks,
Pete
I think, you can use two relatively cheap cameras, one for insects, one for birds and mammals, placing them at appropriate distances from your baits. We were using Genius E-face 1325 R to monitor Myrmica rubra behavior. It works at night two.
Imagery that allows you to monitor a broad view of your bait stations is unlikely to support identification of tiny insect species. Just increasing the pixel resolution will not help much.
Perhaps better to tackle the task in two stages. Continuous video as you outlined will let you to assess the range of predators, whether diurnal or nocturnal, how quickly the bait is consumed, etc. Then you can decide what kindt of detail will be most useful for your research questions.
In the next stage you'll be able to choose an optimal camera and position it according to purpose. E.g. you need fast response if your main predators are birds that flit in an grab the bat, you need a very close position if the main predators are tiny insects, you might want to position the camera side-on (rather than overhead) if some of your birds/mammals are hard to ID from directly above, and so on.
Good luck with the project!
Thanks for your answers Julia, Marianna and Paul, all very useful
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology