A lightweight steel surface defect detection method, device and processing equipment
By improving the C3k2MFB structure of the YOLOv11 model and introducing multiple modules, the problems of low efficiency and insufficient accuracy in traditional detection methods have been solved, achieving efficient and lightweight detection of steel surface defects.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- WENHUA UNIV
- Filing Date
- 2026-04-23
- Publication Date
- 2026-07-10
AI Technical Summary
Traditional steel surface defect detection relies on manual operation, which is inefficient and costly. Furthermore, existing computer vision methods struggle to achieve both high detection accuracy and lightweight design, leading to missed detections and false detections.
Based on the YOLOv11 model, the backbone and neck network are reconstructed through the C3k2MFB structure. The MultiScaleFeatureMixer and DynamicSepConv2d modules are combined to enhance the feature representation capability. The adaptability and detection accuracy of the model are improved through linear deformable convolution and Triplet Attention modules.
While ensuring high detection accuracy, it effectively reduces the number of parameters and computational load, has good generalization ability, and is suitable for lightweight detection in industrial terminals.
Smart Images

Figure CN122368626A_ABST