Skip to main content
Back to search results

AI enthusiast to design and build a new AI based animal welfare support system

Uninhibited Ducks

Do you love animals and also technology?

Are you excited about the ways that AI can aid those of us who work in the animal rescue field, to learn more about the animals we help? And what about if something could be designed with AI and a little electrical/electronics skills to not just track or monitor, but actively engage with birds in a situation where a bird is being harmed, and assist that bird, by triggering a strobe light which will confuse and discourage the attacker?

A North Canterbury bird rescue, founded and operated by a person who works in the technology field is looking for someone super clever, who is a bit of a jack of all trades (or it could be a group of clever people with various skills including hardware, electronics, hardware programming, and AI) to investigate ways to utilise AI, plus cameras to identify at night, unwanted behaviour in birds, or predatory behaviour, and respond to this behaviour in two ways:

1: Recording the behaviour on wildlife or night camera, so myself and you/your team can learn what actually happened and when the software was triggering appropriately and when it was not.

2: Responding to the behaviour by turning on a light on a strobe setting. And ideally, designing something which can pivot to aim the strobe light at the exact direction of the triggering event. Nice to haves would also be time stamping, battery monitoring, alerts to low battery, and any monitoring to assist in identifying areas of inefficiency (ie false triggers wasting battery capacity, obstructions to camera view, etc) .

I have already discovered that a strobe light at night has a harmless, yet confusing and disconcerting affect on birds (predators is to be discovered) and this is a potential way to prevent unwanted behaviour towards other birds during the mating season at night.

I am really interested in learning more about how well this can work, particularly if it could be engineered to trigger remotely, based on AI learning through images and video of the behaviour we want to trigger the strobe lights.

I have an IT background but I do not have the time to acquire the skills in AI programming, plus the hardware selection and HW dev, to design and build something like this.

One question up front I need someone to answer is whether wildlife cam type footage is going to be high enough quality to train and trigger something like this at night? (Daytime footage of trigger events can be recorded but it is during both day and at night I would like this system to be able to detect and react to trigger events)

I am hoping there might be some skilled people out there (maybe students in their final year looking for an interesting and exciting project/thesis topic, maybe IT professionals who have a little time on their hands in this current recession market).

You need to be willing to investigate appropriate hardware (on a budget) AI technologies, and all other design and build tech, from coding/dev, to design, build, right down to soldering the thing together!

As the rescue is based in Nth Canterbury, it would be helpful if you are close enough to be able to visit to see what we do, and also to test any prototypes. However for the right person(s) this could be worked remotely, though some more stringent background checks would be needed then mostly due to the hardware needing to be sent to you.