Burrowing Owl Behavior Classification

Burrowing owls (Athene cunicularia) are just one of the many animals that we are seeing a decline in their population. This is due to pesticides, habitat destruction, climate change, and many other factors affecting their way of life. We plan to help these owls by getting a better understanding of their ecology. We are assisting the San Diego Zoo Institute for Conservation Research to design technology that would be able to help identify and classify their behavior from images.

Using labeled and unlabeled data, we are planning to use machine learning methods to help identify characteristics from taken images eliminating the stage for the researchers to manually identify and describe the details within the images. The ability to identify the behavior of the owls will help researchers automate the recording and labeling of their data in a more efficient manner.

In the first stage of this project, we are testing multiple software to produce the best results when it comes to the detecting the burrowing owls and classifying their species. Once we know the results, we can decide on the best option to properly identify the contents of the images with high accuracy. We can use this to help filter out the images for images that could potentially hold valuable information regarding the owl’s behavior.

The second stage of the project is to design methods to analyze the images to determine the owl’s behavior. We are considering multiple options for this stage of the project, but currently, we are making detection and species classification a priority.

Overall, this project is still in the early stages of development so this is still a very open-ended project when it comes to workflow and design. Feel free to contact us through Slack for any questions related to this project.

Slack Channel: #burrowing-owl

Project Lead: Nathan Hui (nthui@ucsd.edu)

In the video above we show some of the progress made during the summer of 2020.