摘要
依据深度学习算法可以自主进行特征学习和识别的特点,提出一种基于主成分分析法的深度学习模型,构建印刷标签检测系统,从中进行信息提取和学习处理。实验结果表明,该方法的分类识别率高达94.6%。与传统标签缺陷检测算法相比,这种研究方式更加注重原始图像的特征提取,无需进行复杂的模板制作,实现方法简单,适应性强。
According to the characteristics of autonomous feature learning and recognition of deep learning algorithm,a deep learning model based on principal component analysis method is proposed,and a printing label detection system is constructed for information extraction and learning processing.The experimental results show that the classification recognition rate of the method can reach up to 94.6%.In comparison with the traditional label defect detection algorithm,this research method pays more attention to the feature extraction of the original image,needn′t fabricate the complex template,and has simple implementation method and strong adaptability.
作者
李致金
李培秀
朱超
LI Zhijin;LI Peixiu;ZHU Chao(School of Electronic and Information Engineering,Nanjing University of Information Science and Technology,Nanjing 210044,China)
出处
《现代电子技术》
北大核心
2019年第7期153-156,共4页
Modern Electronics Technique
基金
国家自然科学基金(41575155)~~
关键词
机器视觉
深度学习
主成分分析法
标签缺陷
人工智能
模式识别
图像分类
machine vision
deep learning
principal component analysis method
label defect
artificial intelligence
pattern recognition
image classification