一种基于图像展开与拼接的精密轴承表面缺陷光学检测方法

Optical Test Method for Precise Bearing Surface Defects Based on Image Expanding and Splicing

建立一种基于机器视觉的精密轴承表面缺陷光学检测系统。利用图像展开和拼接技术获得轴承侧面完整而又没有重复的二维图像,在此基础上对微小轴承表面缺陷进行检测、缺陷提取和分类。实验结果表明:采用该方法能够快速、高效地检测出微小精密轴承表面大于10μm的缺陷形貌;能够准确地对凹坑、裂纹和划痕缺陷进行分类。

A kind of precision bearing surface defects optical test system was established based on machine vision. Image expanding and splicing technology was used to acquire the two dimensional （2-DOF） image of bearing with whole profile and no repeat, on this basis, the bearing surface tiny defects detection, defects extraction and classification were carried out. Experimental results show that using this method can quickly and efficiently detect the precision bearing surface tiny defects which are greater than 10 μm; and can accurately classify pits, cracks and scratches of defects.