基于机器视觉和图像处理的包装印刷缺陷检测方法

Method of Packaging Printing Defects Detection Based on Machine Vision and Image Processing

为提高包装印刷品缺陷的识别率,基于机器视觉设计了一种印刷缺陷检测方法。分析了常见缺陷类型,利用数字信号处理器TMS320DM642搭建了检测系统同时论述了包装缺陷检测流程。对传统小波变换进行了改进,可增强图像特征信息、提高识别率。重点给出了几种缺陷特征提取方法,包括圆形度、长宽比、灰度标准差等。通过仿真实验验证检测方法的准确性。仿真结果表明:本文所述算法和系统能够显著提高包装印刷缺陷检测准确度,可以达到99%。机器视觉完全满足实时性、高识别率等要求,可提高包装效率以及印刷材料利用率。

In order to improve the recognition rate of package printing defects, a printing defect detection method is designed based on machine vision. The common defect type is analyzed. The digital signal processor TMS320 DM642 is used to set up the detection system and the packaging defect detection process is also discussed. The traditional wavelet transform is improved to enhance image feature information and improve recognition rate. The extraction methods of defect features include circularity degree; length-width ratio and gray standard error are given. The accuracy of test method is verified by simulation experiment. The simulation results show that the algorithm and system described can improve the accuracy of packaging printing defects, which can reach 99%. The machine vision fully meets the requirement of real-time and high recognition rate, which can improve the packaging efficiency and the utilization of printed materials.