| 摘 要: 由于在带钢的生产过程中会出现多种表面缺陷，因此本文中研究了一种基于图像处理的带钢表面缺陷检 测改进算法对表面缺陷进行有效检测。算法中对边缘检测、图像分块、连通域分析等过程进行了改进，并提出了一种针 对带钢图像的图像二值化算法，相较于传统的缺陷检测算法，本文中的检测算法在保证处理速度的同时，可以使处理的 图像细节更完整清晰，缺陷定位更准确，且总体的检测正确率在90％以上，为后续缺陷分类提供更加准确的数据支持， 可有效解决带钢表面缺陷检测问题，对企业生产过程中的技术改善起到至关重要的作用。 |
| 关键词: 带钢；缺陷检测；图像处理 |
|
中图分类号: TP391
文献标识码: A
|
|
| Research on the Detection Algorithm of Strip Steel Surface Defects Based on Image Processing |
|
SUN Guangmin,LIU Peng,LI Zibo
|
( Faculty of Information Technology, Beijing University of Technology, Beijing 100124, China)
|
| Abstract: Since there are many kinds of surface defects in the process of strip production,an improved algorithm based on image processing is proposed to detect the surface defects effectively.In the algorithm,the process of edge detection,image patching,and connected domain analysis are improved,and a binaryzation for images of strip steel is proposed.Compared with the traditional defect detection algorithm,this algorithm,in which the processing speed is guaranteed,can at the same time make the images clearer,the defects more detailed,the positioning more accurate,with over 90% of correct detection rate,which can provide more accurate data support for the subsequent classification of defects and effectively solve the problem of surface defect detection of steel strip,thus playing an important role in the technology improvement in the process of manufacturing. |
| Keywords: strip steel;defect detection;image processing |
