本文是针对于应用在精整线上冷轧带钢表面质量自动检测系统的研究, 该系统用于在线检测划痕、

折印、锈斑、辊印等常见的冷轧带钢表面缺陷。系统采用多个面阵摄像头同步采集图像的方式, 并通

过由光源构成的“ 暗场照明” 方式实现对缺陷的检测。系统采用了并行计算机制和多线程机制, 并对

算法进行了优化, 从而满足数据的实时处理要求。-This article is for applications in the finishing line in the cold rolled strip surface quality automatic detection system, the system for online detection of scratches, off India, rust, India and other common cold roll steel surface defects. Array system uses multiple cameras simultaneously capture images the way, and by the light sources of the " dark-field illumination"　method to achieve defect detection. System uses a parallel computer system and multi-threading mechanism, and the algorithm is optimized to meet real-time data processing requirements.