Computer Vision for species detection on transportation projects
Focus Area
Wildlife & Ecosystems
Subcommittee
Natural Resources
Status
Archived
Cost
$100k-$249k
Timeframe
1-2 years
Research Idea Scope
Computer Vision is a subfield of Artificial Intelligence (AI)/Machine Learning (ML) that uses image recognition to identify wildlife species. The technology has multiple applications for transportation projects including protected species detection and assess of wildlife use of mitigation measures such as crossing structures. A Computer Vision model would be developed utilizing camera trap data of target species and deployed to detect species on transportation projects using camera traps.
Urgency and Payoff
Camera traps often capture 1000s of images of wildlife that require identification to the species level. Computer Vision automates image recognition and has the potential for significant cost savings compared to traditional, manual image recognition. Computer Vision can also be deployed for passive species identification that is less detrimental to sensitive species compared to traditional presence/absence survey methods.
Submitted
06/24/2022