摘要
针对植物纤维形态特征识别中传统方法效率低、精度不足的问题,分析现有纤维研究的现状,通过计算机图像处理技术结合MATLAB编程实现对马尾松、麦草、竹和桉木四种植物纤维的形态特征提取与鉴别。方法包括纤维样品制备、图像采集与处理、特征值选取及程序设计,重点研究纤维长度与宽度的测量以及灰度和灰度方差的计算。结果表明,基于纤维宽度、灰度及其方差的特征值选取方法能够有效区分不同植物纤维,结合图像处理技术的鉴别程序准确率达到80%以上,为植物纤维的快速识别提供了科学依据。
Aiming at the problem of low efficiency and insufficient accuracy of traditional methods in plant fiber morphological feature identification,the present situation of fiber research is analyzed,and the morphological features of Pinus massoniana,wheat straw,bamboo and eucalyptus are extracted and identified by computer image processing technology combined with MATLAB programming.Methods include fiber sample preparation,image acquisition and processing,feature value selection and program design,focusing on the measurement of fiber length and width and the calculation of gray scale and gray variance.The results show that the feature value selection method based on fiber width,gray scale and its variance can effectively distinguish different plant fibers,and the accuracy of the identification program combined with image processing technology is over 80%,which provides a scientific basis for rapid identification of plant fibers.
作者
杨佩
尤国强
YANG Pei;YOU Guoqiang(Xi’an Fanyi University,Xi’an 710105,China)
出处
《造纸科学与技术》
2025年第10期54-57,共4页
Paper Science And Technology
基金
人工智能翻译陕西省高校工程研究中心(2023SGCZX01)
西安翻译学院2024年度科研机构专项项目(2024Z22)
西安翻译学院科研团队(2023SKXJCTD02)。
关键词
计算机
图像处理技术
植物纤维
形态识别
噪声处理
computer
image processing technology
plant fibers
morphological recognition
noise treatment
