Field evaluation of an agricultural weed detector using <scp>YOLO</scp> image recognition: background conditions affect detection performance
This study highlights the importance of background in image-based weed detection and identifies limitations in detector applicability. Our findings are expected to support more-efficient development and underscore the importance of sharing applicability data for improved weed detection tools. © 2025 Society of Chemical Industry.
Authors
- Nozomi Ihara
- Saeko Matsuhashi (ORCID: https://orcid.org/0000-0002-8510-3599)
- Kozue Sasaki
- Akira Koarai
- Yuki Akamatsu (ORCID: https://orcid.org/0000-0002-1519-2805)
- Namiko Yoshino (ORCID: https://orcid.org/0000-0002-9685-040X)
- Hidenori Asami (ORCID: https://orcid.org/0000-0001-9660-4068)
- Ryo Sugiura (ORCID: https://orcid.org/0000-0003-4156-8025)
- Motoaki Asai
- Yusuke Kowata
Institutions
- Institute of Livestock and Grassland Science (JP)
- Research Center for Agricultural Information Technology (JP)
- Western Region Agricultural Research Center (JP)
Publication Details
- Journal
- Pest Management Science
- Published
- 2025-06-18
- DOI
- https://doi.org/10.1002/ps.70009
- Citations
- 1
- Primary Topic
- Smart Agriculture and AI
- Type
- article
- Field-Weighted Citation Impact
- 1.75