An improved YOLO11 network integrated with sample generation for the intelligent detection of steel plate surface defects
Luya Yang, Min Zhang, Yaxian Gao · 2026
Read An improved YOLO11 network integrated with sample generation for the intelligent detection of steel plate surface defects by Luya Yang, Min Zhang, Yaxian Gao (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/plos-pone-0356877.txt.
Description
Luya Yang, Min Zhang, and Yaxian Gao built a three-part system for spotting defects on steel plate surfaces in real time. Flaws hide in low-contrast images on the factory floor. Genuine defect photos are also scarce. Their 2026 study attacks both problems. An adaptive step boosts contrast by merging color channels by information content. A generative network then synthesizes extra defect examples for training.
The detector itself is a slimmed-down version of a current YOLO model. It uses ghost convolutions and attention to stay fast. Together the pipeline lifted detection accuracy by several points. It cut parameters by about a fifth. It reached strong scores at a few milliseconds per image. Steel makers could run such a system on the line without slowing production.
Catalog record
| CREATOR | Luya Yang, Min Zhang, Yaxian Gao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 60844 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |