一种钢球表面缺陷光学检测新方法

A New Method of Optical Detection for Steel Ball Surface Default

为降低企业生产成本,简化钢球检测系统,提出了一种基于机器视觉技术钢球表面缺陷光学检测新方法,采用双目CCD传感器和平面镜反射成像同时获取上下左右四个表面的钢球表面图像,经过图像处理后进行缺陷分级,以取代传统的钢球表面展开机械装置。此外,在照明装置上,设计了一种基于LED面阵光源和磨砂玻璃的方法来模拟均匀散射光线,以提高成像质量。实践表明使用本文方法能够实现钢球表面缺陷的自动检测,简便高效,具有较高的可靠性。

To reduce production costs, simplify the steel ball detection system, a new method of steel ball surface default detection based on machine vision technology was presented. The method, which using two CCD sensors to capture top and bottom steel ball images separately, while adopt two flat mirrors to capture the sides steel ball images based on reflection imaging. Moreover, a new illumination device consisting LED arrays and frosted glass was designed to construct uniform light. By experiment, this method can implements automatic detection of steel ball surface fault simply and effectively.