带钢表面缺陷检测方法研究进展

Research progress on surface defect detection methods of strip steel

带钢是一种重要的钢铁材料,工业生产过程中带钢表面会产生各种不同的缺陷。带钢表面缺陷对产品有重要影响,其特征复杂、多样且不易获取,因此带钢表面缺陷检测一直是研究的重点内容。对带钢表面缺陷检测技术方法的研究进展进行了论述与分析。结合带钢表面缺陷种类,对传统的带钢表面检测方法如人工检测、红外检测、涡流检测和漏磁检测等优缺点进行比较分析,得出这些方法存在检测速度低、无法达到实时在线检测和需要人为干预等缺点。最后对机器视觉的检测方法开展了归纳总结,对基于深度学习的机器视觉识别表面缺陷的原理和方法进行了详述并对未来发展趋势进行了展望。

Strip steel is an important steel material,however,there are various defects on strip steel surface in the process of industrial production.Strip surface defects have an important impact on products,and their characteristics are complex,diverse and difficult to obtain.Therefore,strip surface defects detection has always been the focus of research.The research progress of strip surface defect detection methods were discussed and analyzed.The advantages and disadvantages of traditional strip surface detection methods were compared and analyzed combined with the types of strip surface defects,such as manual detection,infrared detection,eddy current detection and magnetic flux leakage detection.These methods have drawbacks such as low detection speed,inability to achieve real-time online detection,and the need for human intervention.Finally,the detection methods of machine vision were summarized.The principles and methods of surface defect vision based on depth learning machine recognition were described in detail,and the future development trend was also prospected.