South KoreaPublic sector2020 – ongoingKorea National Park Service
Deep-learning detection of dead conifers in Jirisan
Korea National Park Service used aerial-image AI to identify 54,781 dead trees in Jirisan
Problem
Surveying dead conifers in steep mountain terrain was slow and dangerous. Some areas were inaccessible to field experts.
Approach
The park service developed deep-learning image analysis with Dabeeo and Samah Aerial Survey. Researchers trained it using roughly 4,000 dead trees. In November 2020, it identified dead trees across approximately 41 square kilometres and compared its detections with expert interpretation.
Results
Takeaway
Detection differed substantially between standing and fallen dead trees. The average accuracy concealed weaker performance on fallen trees.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Computer vision
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
Yonhap relayed the park service's aerial-image survey results. The system identified 54,781 dead trees, with reported average accuracy of 72.9% and lower detection performance for fallen trees.
Full details are in the original sources.