基于BP神经网络的纸张缺陷检测与识别研究

Research on paper defect detection and recognition based on BP neural network

纸张表面缺陷会直接影响印刷产品的质量。为了快速、准确地检测出纸张缺陷,本文提出了一种基于BP神经网络的纸张缺陷检测与识别的方法。先将纸张缺陷经过形态学处理,再进行形状分析,然后把距离、面积、延长因子和圆度因子四个特征参数输入神经网络进行训练,最后利用训练后的神经网络对纸张缺陷类型进行识别。实验表明:将BP神经网络用于纸张缺陷检测中,能有效地检测缺陷类型,并准确识别常见的尘埃、孔洞、裂口和褶子四种纸张缺陷。

The surface of paper defects will directly affect the quality of printing products;to detect paper defects quickly and accurately,a method for paper defects detection and recognition based on the BP neural network is proposed.With the morphological the paper samples treated and its shape analyzed,the four characteristic parameters are input into the neural network for training,and the trained neural network is used to identify the types of paper defects.The experiment shows that the BP neural network can be used for detecting the defects of paper,and that it can effectively defect detection types and identify four kinds of paper defects accurately,such as dust,holes,cracks,and folds.