钢化玻璃缺陷实时检测系统的研究

Study on real-time inspected system of rempered glass defect

提出了一种基于图像处理的玻璃缺陷实时检测系统。阐述了系统的基本原理和结构,着重研究了系统的核心技术:中值滤波、图像分割、边缘检测、缺陷定位、参数计算与缺陷识别。实验结果表明,该系统的检测识别率高,达到95%;速度快,处理周期约350 ms;抗干扰能力强,基本能够实现钢化玻璃缺陷的在线检测要求。

A real-time inspected system of tempered glass defect based on image processing is proposed. Its running principle and hardware structure are elucidated and the nuclear technologies are focused on improving and researching,including self-adaptive filtering,image segment,edge detection,defect positioning,parameters calculation and defect classification. It features high detection precision. Which is up to90%,high efficiency which is about 350 ms one cycles and strong anti-interference ability. It can satisfy the online detected requirement.