基于CCD光谱系统检测纸制品中二次纤维的含量  

Determination of Recycled-Fiber Content in Paper by Optical Sensor

作　　者：辛丽平 周玉国[1] 丁新平[1] XIN Li-Ping1,2, ZHOU Yu-Guo1,DING Xin-Ping1(1. School of Automation Engineering, Qingdao University of Technology, Qingdao 266033, China; 2. State Key Laboratory of Pulp and Paper Engineering, South China University of Technology, Guangzhou 510640, Chin)

机构地区：[1]青岛理工大学自动化工程学院,山东青岛266033 [2]华南理工大学制浆造纸工程国家重点实验室,广东广州510640

出　　处：《造纸科学与技术》2018年第1期43-46,85,共5页Paper Science & Technology

基　　金：国家自然科学基金青年基金项目(21606141);山东省自然科学基金博士基金项目(ZR2016FB04)

摘　　要：提出了一种基于线阵CCD光谱传感器结合单因变量偏最小二乘回归的二次纤维含量检测技术。首先,搭建了基于线阵CCD光谱传感器的二次纤维含量检测的光学传感系统,可现场快速的获得纸制品的特征光谱信号;针对光谱信号存在的干扰问题,引入了SG平滑和SG求导算法对信号进行预处理,并使用相关系数法剔除冗余无关波段;基于特征波段的特征光谱信号建立了二次纤维含量检测的PLS1回归模型。实验研究表明,基于线阵CCD的光谱系统可以准确的预测训练集和测试集样本的二次纤维含量,其相关系数均在0.99以上,相对均方根偏差均小于5%。本研究对于在确保纸制品的质量和使用安全性的同时,提高二次纤维的利用率,进而推进造纸工业低碳经济模式的发展,将具有积极的意义。A method of determining the content of recycled fiber in paper by CCD sensor coupled with Partial Least Square Regression（ PLSR） is proposed.The system equipped with a CCD sensor,some optical fibers and other electronic accessories was set up to collect the spectra of paper samples.To remove the interference,the spectral signal was pre-processed succes sively by the Savitzky-Golay smoothing and the first derivative calculations.The PLS1 model was developed based on the processed spectra.The results showed that PLS1 model performed well for the samples from both the training set and the test set,the R2 were more than 0.99,and the RRMSE were less than 5%.The present method is non-destructive,no sample pretreatment,no consumption of chemicals and reagents,no need of both a qualified laboratory technician and laboratory grade facilities,and particularly fast determination the recycled fibers content in point-of-sale samples from commercial markets.

关 键 词：CCD传感器 光谱信号 PLS1 二次纤维 

分 类 号：TS77[轻工技术与工程—制浆造纸工程]