基于机器视觉的高速带钢孔洞检测系统

针对宝钢冷轧1420酸轧联合机组带钢孔洞缺陷在线检测需求，结合机组现场条件，自主开发了恶劣危险环境下的高速带钢孔洞检测系统。该系统以机器视觉技术为基础，配置高性能透射式照明光源和高速线扫描CCD图像传感器，能够获得高质量的带钢图像；系统检测软件采用快速高效的目标识别和边缘检测算法，可在30ms内完成1幅图像的处理计算，实时判定孔洞的存在与否，并将检测结果发送至服务器。系统采用C/S模式作为网络架构，实现检测数据在客户端和服务器之间的可靠传递。系统两年多的上线运行证明：该系统在冷轧机出口十分恶劣的环境条件下，可长期连续工作，准确高效地检出带钢孔洞缺陷，具有卓越的检测性能和良好的稳定性。

High-speed hole inspection system for steel strips based on machine vision

The high-speed hole detection system for strip steel was developed to meet the urgent needs for detection of hole defects in the strip steel produced by 1420 CDCM mill in Baosteel Co., Ltd. This system, which was designed based on machine vision technology, has been working in the extremely tough conditions for more than 2 years. Configured with transmission illumination and high-speed linear scanning CCD cameras as image sensors, high quality images of steel strip can be obtained in real time, which can be processed at a speed of 30 ms/frame with the detection software. The software transfers the control instructions and hole-detecting results to the server through a C/S network structure that consists of 3 industrial PCs, and a terminal connecting the server computer displays inspection results and sends alarm signals, Due to its state-of-the-art design, high-efficiency object detection and edge detection algorithms, the outstanding performance and good reliability of this system have been fully proved in the practice for more than 2 years.