基于深度学习的宣纸缺陷识别软件设计  

Design of Defect Recognition Software for Rice Paper Based on Deep Learning

作　　者：张志强[1] 谢勇[1] ZHANG Zhiqiang XIE Yong(Xi'an International University,Xi'an 710077,China)

机构地区：[1]西安外事学院,陕西西安710077

出　　处：《造纸科学与技术》2024年第1期98-101,共4页Paper Science & Technology

基　　金：2021年度陕西省高等教育教学改革研究项目(21ZY015);2021年度西安外事学院教育教学改革研究项目(2021B10)。

摘　　要：作为我国优秀文化遗产书法与国画的主要载体,宣纸需始终保持较高质量,最大程度上避免缺陷以影响外观与使用,这就要求造纸企业必须高度重视宣纸缺陷识别与检测,以机器识别软件替代人工识别检测,并引进合适的识别算法。提出了基于深度学习之卷积神经网络的宣纸缺陷识别算法,且由此进行了宣纸缺陷识别软件设计。结果发现,此软件不仅运行速度快,识别率高,而且识别分类准确率可达98.93%,可完全贡献于宣纸质量提高,值得广泛推广与应用。As the main carrier of China's excellent cultural heritage calligraphy and traditional Chinese painting,rice paper needs to always maintain high quality and avoid defects to the greatest extent that may affect its appearance and use.This requires papermaking enterprises to attach great importance to rice paper defect identification and detection,replace manual identification and detection with machine recognition software,and introduce appropriate recognition algorithms.This article proposes a deep learning based convolutional neural network algorithm for rice paper defect recognition,and designs a rice paper defect recognition software based on it.The results show that this software not only runs fast and has a high recognition rate,but also has a recognition and classification accuracy of 98.93%,and it can fully contribute to improving the quality of rice paper and is worthy of widespread promotion and application.

关 键 词：深度学习 卷积神经网络算法 宣纸缺陷 缺陷识别 软件设计 

分 类 号：TS736.4[轻工技术与工程—制浆造纸工程]