Research on Strip Steel Defect Detection Technology Based on Improved YOLO11
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Strip steel is a fundamental industrial material, and the detection of its surface defects is directly related to product quality and quality inspection efficiency. However, existing detection methods often suffer from problems such as missing detection of small defects and inaccurate recognition under complex backgrounds, and it is difficult to balance detection accuracy with the real-time performance required by production lines. To solve these problems, this paper makes targeted improvements to the YOLO11 detection model: first, two key feature compression modules in the model are replaced to preserve the detailed information of small defects to the maximum extent; second, a global feature recognition module is added to enhance the ability to distinguish defects from the background. Experimental verification shows that the improved model not only significantly reduces the missing detection rate of small strip steel defects and achieves better detection accuracy than the traditional YOLO11 model and other mainstream detection methods, but also demonstrates excellent robustness through the interference test set. Meanwhile, it maintains a fast detection speed and can meet the real-time detection requirements of industrial production lines without consuming a large amount of computing resources, thus facilitating deployment on factory equipment. This research provides a feasible solution for the efficient and intelligent detection of strip steel surface defects, and has important significance for promoting the intelligent transformation of industrial quality inspection.
Fields of Study & Subjects
Industrial Vision Systems and Defect Detection
Non-Destructive Testing Techniques
Advanced Neural Network Applications
Publication details
- Journal / source
- International Conference on Communication Networks and Machine Learning (CNML)
- Pages
- 726–730
- Publication date
- 30 January 2026
- Language
- EN