United StatesPublic sector2019 – ongoingNational Audubon Society
AI waterbird counts from drone imagery
Audubon reported almost 90% accuracy in bird-image tests
Problem
Remote colonies were difficult to survey.
Approach
Audubon trained an image classifier on labeled bird photographs. It tested identification and counts using drone images.
Results
almost 90%test accuracy reported by Audubon
Takeaway
The tests used four drone images, and identification was more accurate for Brown Pelicans than for terns or Black Skimmers.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
[1]National Audubon Society14 Mar 2019
Audubon reported a completed bird-counting pilot. Pelican identification performed better than tern and skimmer identification.
Full details are in the original sources.
Added 1 Nov 2026 · Updated 1 Nov 2026