刘晓杰+罗印升+张旻+范洪辉

摘 要： 针对零部件表面缺陷检测精度问题，提出一种基于机器视觉的零部件表面缺陷检测方法。传统的利用机器视觉对零部件表面缺陷检测方法中，由于零部件表面的光学反射特性，因此无法对零部件表面缺陷进行高精度的检测。提出的基于机器视觉的零部件表面缺陷检测方法引进了差影法检测模型，根据部件表面特征，利用分段线性灰度算法对部件表面细小的缺陷进行区域检测，并且结合了灰度共生矩阵的换算熵作为判定的依据，最终建立的缺陷检测模型是利用矩阵方位度和相似度之比进行高精度的检测。为了验证设计的基于机器视觉的零部件表面缺陷检测方法的有效性，通过仿真试验证明了该设计方法，结果表明该方法能够有效地解决零部件表面缺陷檢测的精度问题。

关键词： 机器视觉； 零部件表面缺陷； 差影法； 灰度计算

中图分类号： TN912.205?34； G420 文献标识码： A 文章编号： 1004?373X（2017）24?0181?03

Abstract： In allusion to the precision problem of parts surface defect detection， a machine vision based detection method for defects on parts surface is proposed. For the traditional parts surface defect detection method based on machine vision， it is not possible to perform high?precision parts surface defect detection due to the optical reflection characteristic of parts surface. The difference image detection model is introduced in the proposed parts surface defect detection method based on machine vision. According to the features of parts surface， the piecewise linear gray algorithm is adopted to perform area detection for small defects on parts surface， and with the conversion entropy of the gray level co?occurrence matrix as the judging basis， the finally?constructed defect detection model utilizes the ratio of matrix orientation degree to similarity degree to perform high?precision detection. In order to verify the validation of the designed parts surface defect detection method based on machine vision， the simulation experiment was carried out. The experimental results show that the designed method can effectively resolve the precision problem of parts surface defect detection.

Keywords： machine vision； parts surface defect； difference image method； gray calculation

伴随着工业的不断发展，各领域中使用的零部件都要经过多重检测才能进行使用，由于零部件生产中都是大批量生产，致使对于零部件的检测要求快速的同时还要保持高精度[1?2]。零部件在生产加工过程中，由于使用材质以及加工成型工艺的问题，会导致零部件表面出现裂纹、划痕、针孔等多种类型的缺陷[3]。这些缺陷会极大地影响工件的美观程度，同时降低工件的使用性能，严重的可能会导致一定的安全隐患。……