Artificial Intelligence to Help Intelligent Dolphins
We’ve developed an AI powered tracking UAV (unmanned aerial vehicle) that will autonomously find and follow Māui and Hector’s dolphins, and uniquely identify them. Once we know where they are and where they travel, we can protect our taonga from extinction. We won’t stand by as the world’s rarest dolphin population drops below 63.
The Māui Project
The biggest threats are location related. By identifying individual dolphins, knowing where they are and when they are there, we can help protect them from falling victim to fishing, predators, climate change, toxoplasmosis and other mishaps. It will help tremendously to halt the decline towards extinction.
MAUI63 helps to relieve all the main threats to our Māui and Hector’s dolphins. The project in a nutshell:
– Uniquely identify individuals via fin markings – Distinguishing between adults and calves – Find and track dolphins via surveys, to create accurate spatial distribution models – Reveal temporal changes in dolphin movement patterns – Use the data to inform threat risk models informed by natural (e.g. shark predation) and anthropogenic factors (e.g. fisheries, climate change, toxoplasmosis).
By automating the surveying process, regular location data of critically endangered marine animals can finally be captured!
No More Fishing Accidents
Commercial fishing is the largest anthropogenic threat to the dolphins, but also an economically important industry to New Zealand. Restricted zones and sanctuaries are in place, but without continuous monitoring they are hard to manage, and the sanctuaries can’t be extended appropriately.
Our up to the minute data will allow policy makers to accurately update the restricted zones, ensuring fishing boats are nowhere near the dolphins!
We have built an AI object detection model that can uniquely identify Māui and Hector’s dolphins from aerial footage. Our large UAV will utilise the AI and fly at 120km/h looking for dolphins. Once it has spotted them, it will automatically circle and follow them for up to 6 hours.
We are continuously improving the AI. What this means is that in the very near future we want to uniquely identify dolphins, based on their fins, and distinguish calves from adults. We want to be able to tell exactly which dolphin has been where and when! This data is extremely useful for research into declining populations and deaths.
Once we have collected a significant amount of data, we plan to implement predictive models and further utilise Machine learning to predict location behaviour among other data between flights, resulting in a reduced flight time & overall cost. All this amazing data will be collated and made publicly available to researchers, policy makers and industry.
Follow our progress
Check out the timeline below to see our goals and how we are progressing. You can see exactly where any money raised will go!
MAUI63 – That Name…
Why are we named MAUI63? It’s not just the name of the world’s rarest dolphin. It’s also an acronym for Marine Animal Unmanned Identification, and the 63 represents the count of Māui dolphins when we started.
The name aligns with our goal – to support the growth of the Māui dolphin population – way beyond 63!
Currently we are a small team. But we’re growing and we’re keen to find people to help and support us. MAUI63 was founded by Rochelle Constantine, Willy Wang & Tane van der Boon.
Rochelle is our research lead and marine science extraordinaire. She is an Associate Professor at the University of Auckland with over 25 years of research experience as programme lead on large international and New Zealand based whale and dolphin research projects, including Māui dolphins.
Our operations lead Willy is a type 102 qualified drone enthusiast, registered doctor and is currently completing a PhD in medical science at the University of Auckland. Some of his other work includes biosensor development utilising AI to guide clinical decisions.
Tane is the project and technical lead. His background is in enterprise IT and innovation and has worked on similar projects that included using object recognition AI to keep staff safe from harm within the work environment. Tane is also a type 102 qualified drone pilot and has a passion for technology, nature & wildlife.