WeedCAM: An edge-computing camera system for multi-species weed detection in sugar beet production fields
Computers and Electronics in Agriculture · 2026-03 · 244 · 111498
DOI: 10.1016/j.compag.2026.111498 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computers and Electronics in Agriculture, 2026-03. Verfasst von Zonglin Yang und 8 weiteren Autor:innen. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 55 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Science Foundation Directorate for Technology Innovation and Partnerships, USDA NIFA, NSF.
Abstract (Originalsprache)
This study introduces WeedCAM, a low-cost, near real-time, edge-computing camera system for multi-species weed detection, built on a Raspberry Pi 5 and integrated with a GPS module and LoRa board for geolocation and data transmission. A three-phase framework, including data acquisition, model fine-tuning, and deployment, is proposed to implement detection models on WeedCAM. A total of 5,734 high-resolution (4K) images were automatically collected using WeedCAM, producing a dataset with a long-tail distribution that poses challenges for model training. To address this, Repeat Factor Sampling and Focal Loss were applied during the fine-tuning. Seven object detection models were evaluated, including YOLOX-S, YOLOX-L, Faster R-CNN, Cascade R-CNN, Deformable-DETR, and DINO series. Finally, WeedCAMs with embedded trained models were deployed in the field on pivot-mounted and ground-based installations, detecting weeds at 30-minute intervals and transmitting results to a customized gateway via LoRa. The gateway parsed and mapped these results to our custom-designed website for visualization. Our best model, DINO-Swin/L, set the performance benchmark with a 76.0 overall mAP (IoU=0.5) and strong per-species scores for kochia (76.4 mAP), Palmer amaranth (77.3 mAP), and volunteer corn (75.9 mAP) in 4K resolution image. Despite this, YOLOX-L was deployed on the WeedCAM, as its efficient 8 minute processing cycle represented the better trade-off between accuracy and speed. Field evaluation confirmed that WeedCAM effectively identified weed species and quantities under varying lighting conditions, camera angles, and soil moisture levels during irrigation events. These results demonstrate the practicality of deploying WeedCAM edge-computing deep learning systems for near real-time, multi-species weed detection under sugar beet production fields.
Förderung
- National Science Foundation Directorate for Technology Innovation and Partnerships
- USDA NIFA
- NSF
- Established Program to Stimulate Competitive Research
Zitieren
APA
Yang, Z., Liang, W., Lawrence, N., Qiao, X., Riggan, B., Harveson, R., Chiang, C., Oboamah, J., & Andjawo, D. P. (2026). WeedCAM: An edge-computing camera system for multi-species weed detection in sugar beet production fields. Computers and Electronics in Agriculture, 244, 111498. https://doi.org/10.1016/j.compag.2026.111498
Vancouver
Yang Z, Liang W, Lawrence N, Qiao X, Riggan B, Harveson R, et al. WeedCAM: An edge-computing camera system for multi-species weed detection in sugar beet production fields. Computers and Electronics in Agriculture. 2026;244:111498. doi:10.1016/j.compag.2026.111498
BibTeX
@article{Yang2026,
author = {Yang, Zonglin and Liang, Wei-Zhen and Lawrence, Nevin and Qiao, Xin and Riggan, Benjamin and Harveson, Robert and Chiang, Chi-En and Oboamah, Joseph and Andjawo, Diwenitissiou Philipine},
title = {WeedCAM: An edge-computing camera system for multi-species weed detection in sugar beet production fields},
journal = {Computers and Electronics in Agriculture},
year = {2026},
volume = {244},
pages = {111498},
doi = {10.1016/j.compag.2026.111498}
}
EUVIMED