Case Study: Vale achieves 82% precision on aerial wildlife surveys with LightlyTrain

A Lightly Case Study

Preview of the Vale Case Study

Vale reaches 82% precision on herd size estimation with Lightly

Vale, a wildlife technology company, builds monitoring ecosystems but faced significant challenges scaling its computer vision models. Their field teams generated millions of aerial and trail camera images, far more than could be manually reviewed, and they lacked enough labeled data to train accurate wildlife detectors. Vale turned to Lightly and its LightlyTrain service to solve this scarce data problem.

Lightly implemented a solution using LightlyTrain to build DINOv3-based detectors that integrated directly into Vale's existing pipeline. By training on a combination of synthetic data and a small set of real labels, Vale achieved roughly 82% precision on herd size estimation. This solution from Lighty provided the foundation for automated analysis that would have taken biologists an estimated 16,000 hours to complete manually.


View this case study…

Lightly

10 Case Studies