基于图像处理和BP网络的烟包印刷质量检测

为解决烟包印刷质量在线检测问题,本文提出了一种基于图像处理和BP神经网络的印刷质量检测和诊断方法.利用待测印刷图像和标准模版图像间差异的特征参数,BP神经网络可将印刷品分成优秀、合格以及不合格三类.实验结果表明,现有系统能够满足烟包印刷质量检测的需求,并且能够运用于烟包印刷质量的在线检测中.

Print Quality Detection of Cigarette Wrapper Based on Image Processing and BP Network

To resolve the problem of the print quality detection of cigarette wrapper online, the paper has proposed a method to detect and diagnose the print quality based on image processing and BP network. Using the feature parameters of the subtraction image between the detected print image and the standard template image, a BP network is applied to classify the prints into excellent, qualified and unqualified products. The experiment results have shown that the developed system can meet the demands of the print quality detection of cigarette wrapper and can be used to implement the onlin eprint quality detection of cigarette wrapper.