摘要
随着智能制造技术的持续推进,传统的纸张质量检测手段已无法契合高效、精准且自动化的工业生产所需。基于此,站在智能化角度,深度探究深度学习算法在纸张缺陷检测系统中的优化运用。通过搭建改良后的卷积神经网络(CNN)模型,同时融合注意力机制以及数据增强策略,极大地提高了系统针对多种缺陷(像破洞、褶皱、污点这类)的识别精准度与稳定性。实验得出,优化过的系统在诸多真实场景中,泛化能力和检测速度良好,给纸张生产线的智能质量管控指明了可行性方向。
With the continuous advancement of intelligent manufacturing technology,traditional paper quality detection methods can no longer meet the needs of efficient,accurate and automated industrial production.In this paper,from the perspective of intelligence,the optimal application of deep learning algorithm in paper defect detection system is deeply explored.By building an improved convolutional neural network(CNN)model and combining attention mechanism and data enhancement strategy,the recognition accuracy and stability of the system for various defects(such as holes,folds and stains)are greatly improved.Experiments show that the optimized system has good generalization ability and detection speed in many real scenes,which points out the feasible direction for intelligent quality control of paper production line.
作者
党佳奇
王婉星
DANG Jiaqi;WANG Wanxing(Shaanxi Polytechnic University,Xianyang 712000,China)
出处
《造纸科学与技术》
2025年第9期96-99,共4页
Paper Science And Technology
基金
陕西省“十四五”教育科学规划2024年度课题(SGH24Q521)。
关键词
纸张缺陷
深度学习
卷积神经网络
智能检测
工业视觉
paper defects
deep learning
convolutional neural network
intelligent detection
industrial vision
