基于计算机图像处理的纸浆纤维检测与分类研究  

Research on Detection and Classification of Pulp Fibers Based on Computer Image Processing

作　　者：王娟[1] 张娜[1] 雷虎[1] WANG Juan;ZHANG Na;LEI Hu(Xi’an Fanyi University,Xi’an 710105,China)

机构地区：[1]西安翻译学院,陕西西安710105

出　　处：《造纸科学与技术》2025年第1期25-28,共4页Paper Science And Technology

基　　金：陕西省自然科学基础研究项目(2024JC-YBMS-571)。

摘　　要：纸浆纤维的形态参数和纯净度等决定着纸张的质量与性能。对纸浆纤维进行精准检测和分类是造纸生产中的重要环节,也是保障成纸质量最直接的手段。计算机图像处理技术是近些年随着信息产业的迅速发展而发展起来的新兴分析技术,在纤维检测领域有着明显优势与良好前景。基于此,简述了计算机图像处理技术在纸浆纤维检测中的应用,介绍了纸浆纤维图像预处理过程与算法,确定了对纸浆纤维图像要提取的形态特征和灰度特征,并提出SVM结合卷积神经网络的纸浆纤维分类算法。该方法对纸浆纤维的分类达到了较高的水准。The shape parameters and purity of pulp fiber determine the quality and performance of paper.The accurate detection and classification of pulp fiber is an important link in papermaking production,and also the most direct means to ensure the quality of paper.Computer image processing technology is a new analytical technology developed with the rapid development of information industry in recent years,and has obvious advantages and good prospects in the field of fiber detection.Based on this,this paper briefly describes the application of computer image processing technology in pulp fiber detection,introduces the pretreatment process and algorithm of pulp fiber image,determines the morphological features and gray features to be extracted from pulp fiber images,and proposes the SVM combined with convolutional neural network pulp fiber classification algorithm,which has reached a high level of pulp fiber classification.

关 键 词：计算机图像处理 纸浆纤维 特征提取 SVM 卷积神经网络 

分 类 号：TS72[轻工技术与工程—制浆造纸工程]