摘要
对我国华南地区六种棕榈藤的纤维和导管形态特征进行了观测,研究了应用近红外光谱分析技术预测棕榈藤纤维长度和导管长度的可行性。结果表明：六种棕榈藤的纤维平均长度在1 229~1 917μm,导管的平均长度在1 035~2 129μm;发现棕榈藤纤维长度和导管长度模型的建立采用偏最小二乘法和完全交互验证法,在350~2454 nm谱区内用一阶导数处理的光谱与纤维长度之间建立的模型的相关性较好,校正模型和预测模型的相关系数rc和rp及标准误差SEC和SEP分别为0.98,0.85和70,178,在350~2 500nm全谱区内用一阶导数处理的光谱与导管长度之间建立的模型的相关性也较好,校正模型和预测模型相关系数rc和rp及标准误差SEC和SEP分别为0.97,0.80和101,261。说明红外光谱分析技术可以用来预测棕榈藤的纤维长度及导管长度。
The morphological characteristics of fibers and vessels of six rattan species in Southern China were investigated to study the feasibility of predicting the lengths of fibers and vessels of rattan species with application of analytical technologies of near infrared spectroscopy（NIR）.The results showed that the average lengths of fibers and vessels of six rattan species were from 1 229 to 1 917 mm and from 1 035 to 2 129 mm,respectively.The models of length of fiber and vessel were constructed by combining partial least square（PLS） and full cross-validation,and a good correlation between the length of fibers and the spectrum transformed by the first derivative was found within the spectral range of 350～2 454 nm,and the correlation coefficient（rc and rp） and standard error（SEC and SEP） of calibration model and prediction model are 0.98,0.85 and 70,178 respectively,while a good correlation between the length of vessels and the spectrum transformed by the first derivative was found within the spectral range of 350～2 500 nm,the correlation coefficient（rc and rp） and standard error（SEC and SEP） of calibration and prediction model is 0.97,0.80 and 101,261 respectively.Their model parameters showed that NIR spectroscopic technique can rapidly and accurately predict the lengths of fibers and vessels of the six rattan species.
出处
《光谱学与光谱分析》
SCIE
EI
CAS
CSCD
北大核心
2011年第4期966-969,共4页
Spectroscopy and Spectral Analysis
基金
国家"十一五"科技支撑项目(2006BAD19B01)
国家林业局"948"引进项目(2006-4-96)
国际竹藤网络中心基本科研业务专项基金项目(1632008001)资助
关键词
棕榈藤
纤维长度
导管长度
近红外光谱
Rattan
Fibers length
Vessels length
Near infrared spectroscopy（NIR）
