discussion / AI for Conservation  / 14 April 2024

Has anyone combined flying drone surveys with AI for counting wild herds?

My vision is a drone, the kind that fly a survey pattern. Modern ones have a 61 megapixel camera and LIDAR which is a few millimeter resolution, they are for mapping before a road is built as an example. 

We know where there are large herds, for example wild horses. My larger concern is invasive species, such as hogs. The AI can find and count the animals inside of a large area, using cameras and existing sensors. From there we can take the next action, which is giving contraceptives to the breeding females and slowly reducing numbers. Or fast, if you know where they are. I spoke with Oracle AI and they said that would be easy. Then they asked if I had any funding lol. 

Lars Holst Hansen
Aarhus University
Biologist and Research Technician working with ecosystem monitoring and research at Zackenberg Research Station in Greenland
Conversation starter level 3
Popular level 3
Poster level 2
Reactor level 3
Involvement level 3

Hi Johnathan!

Here are a few examples where UAVs and AI has been used to spot animals. 



A google scholar search as this will find many more:


One thing often forgotten when considering UAVs for aerial surveys like these are that maximum height above ground is normally about 100-120m. This really limits the area one can cover.



That was one of the things I was wondering about, the height that it can resolve animals at. At some resolution it must be able to tell different animals apart. 


My application is for invasive herds, or uncontrolled large animal herds such as wild horses or urban deer. In phase 2 we apply contraceptives to them to humanely reduce numbers.

Our research group has worked on this issue. Here is the paper link: 





Awesome! I am looking at something similar too - also awaiting funding. It's in SE Australia, also focusing on invasive species (red fox), protecting an area that has nesting & migratory shorebirds.

I was looking at infrared images, which will present a new challenge for AI/ML coding as there isn't much by way of training images to work with.

I think a lot depends on whether you are hoping to do the image recognition in real-time (during flight) or after downloading the images.

Also flying transects seems better supported for statistical modelling than camera trap images. 

Very interested to hear how it goes.