RGC-YOLO: a lightweight YOLOv8n-based detector for edge-deployable weed detection in Xinjiang cotton fields
Qianqian Mu, Yongke Li, Nueraili Aierken, Lei Wang, Tao Zhao, Mengli Shan, Bowen Mao, En Yu · 2026
Read RGC-YOLO: a lightweight YOLOv8n-based detector for edge-deployable weed detection in Xinjiang cotton fields by Qianqian Mu, Yongke Li, Nueraili Aierken, Lei Wang, Tao Zhao, Mengli Shan, Bowen Mao, En Yu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13601265.txt.
Description
Qianqian Mu and seven co-authors built a slim weed-spotting model for cotton fields in Xinjiang, in a 2026 paper. Field-side devices need models that are accurate yet light and fast. Their detector builds on a compact version of a popular vision model. Field photos of four weed types formed the data set.
New modules trim the model's size in its deep layers. A fusion block sharpens how features combine. A revised loss function tightens box placement on targets. Scores at strict overlap rose while parameter count fell by about a third. Heat maps showed focus on weeds rather than background. On an edge computing board, the model ran fast in a compact engine. Tests on sesame fields and a second weed data set held up well. The design suits real farm deployment, the authors conclude.
Catalog record
| CREATOR | Qianqian Mu, Yongke Li, Nueraili Aierken, Lei Wang, Tao Zhao, Mengli Shan, Bowen Mao, En Yu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 71 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |