林业工程学报2025，Vol.10Issue(5)：82-90,9.DOI:10.13360/j.issn.2096-1359.202407026
近红外光谱预测芳樟、龙脑樟精油主成分含量
Prediction of principal components of linalool-type and borneol-type essential oils by near-infrared spectroscopy
摘要
Abstract
Using near-infrared spectroscopic analysis techniques,the essential oils and leaves of linalool-type and borneol-type camphor were studied and quantitatively analyzed to facilitate quicker and more convenient determination of major component quantities in these plants.The spectra were scanned individually,and by integrating chemometric techniques with essential oils and leaves as the materials,quick prediction models for the concentration of linalool-type and borneol-type essential oils were developed.The model of linalool-type essential oil,with partial least squares(PLS)computational modeling and Dertivatives(1st BCAP)(db1)preprocessing,performed best in the 1 000.00-1 350.62 and 1 399.78-2 000.00 nm bands,with related coefficient of calibration(RC)and related coefficient of validation(RV)values of 0.962 0 and 0.953 9,respectively,and root mean square error of calibration(RMSEC)and root mean square error of validation(RMSEV)values of 0.011 0 and 0.015 5,respectively.The model of linalool-type leaves performed best in the 1 000.00-2 000.00 nm band,with ds2 preprocessing and PLS algorithm conditions of the built model,with RC and RV values of 0.938 6 and 0.930 3,respectively,and RMSEC and RMSEV values of 0.033 1 and 0.032 1,respectively.Borneol-type essential oil model:using PLS computational modeling,multiple scatter correction(full)(mf),db1,and normalization(to unit length)(nle)three methods combined pre-processing spectra,the constructed prediction model had the best prediction performance;RC and RV values of 0.985 2 and 0.955 8,respectively;RMSEC and RMSEV values of 0.075 9 and 0.078 6,respectively.Borneol type leaves model:choosing principal component regression(PCR)modeling method,1 000.00-1 350.62,1 399.78-2 000.00 nm modeling bands;the best Lobelia prediction model constructed after standard normal variate(SNV)preprocessing;the values of RC and RV were 0.944 2 and 0.953 0,respectively,RMSEC and RMSEV values of 0.112 1 and 0.130 8,respectively.The external verification of the constructed models demonstrated their good performance and the absence of significant differences in the predicted values.All built models performed well when externally verified,with no discernible deviation from the expected values and no significant differences in the results by paired-sample t-test.This suggested that all built models had good predictive ability and can accurately predict the content of the principal components of the essential oils of linalool type and borneol type.This study is the first research related to near infrared spectroscopy(NIR)analysis technology in Cinnamomum camphorar,which provides new ideas and methods for the rapid detection of essential oil principal components in C.camphorar.关键词
近红外光谱/芳樟/龙脑樟/精油/主成分含量Key words
near infrared reflectance spectroscopy/linalool-type/borneol-type/essential oils/main component content分类
轻工纺织引用本文复制引用
涂白连,刘新亮,郑永杰,张月婷,伍艳芳..近红外光谱预测芳樟、龙脑樟精油主成分含量[J].林业工程学报,2025,10(5):82-90,9.基金项目
江西省自然科学基金(20232ACB205018) （20232ACB205018）
江西省林业科技创新专项项目([2021]15号) （[2021]15号）
江西省重点研发计划项目(20202BBF63006). （20202BBF63006）
