摘要： CCL表面的质量标准日益严格,人工检验表面缺陷的传统方法已不能适应市场要求。该文研究了CCL图像的特点,提出了实现CCL 表面质量自动检测的图像处理过程。阐述了利用图像的边缘特征,通过边缘检测、连接和区域填充来分割缺陷的方法,以及一种获取缺陷识 别所需特征值的扫描算法。实验结果表明,利用所提出的图像处理过程,缺陷的漏检率为1％,误检率为3％。

关键词: 缺陷, 边缘检测, 特征值

Abstract: The quality standard for CCL surface is becoming higher and higher,The current inspection method based on operators' vision can't meet market requirement. In this paper, the image processing flow to realize automatic detection is presented, it discusses the segmenting method by means of edge detection, edge join & area fill, and the scanning algorithm to acquire defect features. The experiment results show that the image processing method can make it realize to automatically detect the surface delect of CCL.

中图分类号: 

• TP274.4