作　　者： ; ; ;

机构地区： 华南理工大学机械与汽车工程学院

出　　处： 《测试技术学报》 2010年第6期518-521,共4页

摘　　要： 针对目前企业缺乏钢球表面缺陷检测装备,特别是小直径钢球检测装备的现状,笔者研制开发出一套基于机器视觉、可批量检测小直径钢球表面缺陷的新装备.由特制机电装置将钢球成批自动行列相间地上料并使球面翻滚,相机以一定帧频拍摄多幅钢球表面图像.在保证检测精度和效率的前提下,设计出应用于该装置的图像预处理方法,包括模板标定、图像分割和图像增强.实际检测试验证明,该预处理方法一方面使表面缺陷突显,有利于后期检测,另一方面抑制外接环境引入的各种干扰,消除不必要的影响因素. Aiming at the shortages in the detection means and equipments of steel balls＇ surface defects at present, particularly for small ball inspection, the authors develop a new system based on machine vision for the automatically batch detection of small steel balls. By means of a special mechatronic device the small balls are distributed as an array and loaded in batches automatically, and then they are rolled over so that the camera can takes the images of variational sphere area. In order to improve the detection precision and efficiency, a series of image preprocessing approaches are designed for the system, which includes the techniques of template calibration, the image segmentation and enhancement. The experiment proves that the image preprocessing method on one hand can get the defects obvious in favor of the following detection, on the other hand can inhibit various inferences from external environment and eliminate the needless effect factor.

关 键 词： 表面缺陷 图像预处理 模板标定 图像分割 图像增强

领　　域： [自动化与计算机技术] [自动化与计算机技术]