For university students and researchers, TortugaTracks sits at the intersection of marine conservation, computer vision, and field data collection.

The recognition problem

Detecting and classifying turtle tracks from aerial imagery is a practical computer-vision challenge: variable lighting, shifting sand, tidal washout, and class imbalance between species all make it harder than a clean benchmark dataset. Understanding those constraints is part of building models that work in the field rather than only in the lab.

Methods and data

Robust results depend on careful data collection, annotation, and validation. There is real scope here for work on model architecture, evaluation under domain shift, and the ethics of where and how ecological data is stored and shared.

Collaborate

We’re interested in working with research groups on datasets, methodology, and field validation. If your department works on marine ecology, remote sensing, or applied machine learning, there may be room to collaborate.

Get involved

Researchers can also contribute drone footage and help validate detections through the platform. Details on data access and collaboration will be published here.

This is starter copy — replace it with real methodology, publications, and contact routes.