近红外光谱结合蚁群算法检测花茶花青素含量  

Measurement of anthocyanin content in flowering tea using near infrared spectroscopy combined with ant colony optimization

作　　者：黄晓玮[1] 邹小波[1] 赵杰文[1] 石吉勇[1] 张小磊[1] 

机构地区：[1]江苏大学食品与生物工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2014年第2期165-170,188,共7页Journal of Jiangsu University：Natural Science Edition

基　　金：国家"863"计划项目(2011AA108007);国家自然科学基金资助项目(60901079);全国优秀博士基金资助项目(200968);江苏省农业自主创新项目(CX(11)2028);江苏大学拔尖人才启动基金资助项目

摘　　要：花青素是花茶中的主要质量指标,为了快速准确的检测花茶中花青素的含量,提出一种基于蚁群算法(ACO)结合区间偏最小二乘法(iPLS)的近红外光谱检测方法.原始近红外光谱经过预处理采用ACO-iPLS优选花青素含量对应的特征子区间.当全光谱划分为12个子区间时,ACO-iPLS优选出第1,9,10共3个子区间,在此基础上建立的近红外光谱模型最佳.模型对校正集和预测集相关系数分别为0.901 3和0.864 2;交互验证均方根误差(RMSECV)和预测均方根误差(RMSEP)分别为0.160 0 mg·g-1和0.202 0 mg·g-1.研究结果表明:与常规的iPLS相比,ACO-iPLS不但可以有效选择近红外光谱特征谱区,而且建立的模型具有更高的精度和鲁棒性.To improve the detecting accuracy and rate of anthocyanin content in flowering teas, a detec- tion method was proposed based on near infrared spectroscopy combined with interval PLS and ant colony optimization. The flowering tea spectra were divided into 12 intervals, and three subsets of No. 1, 9 and 10 were selected to establish an iPLS model. The calibration and prediction correlation coefficients were 0.901 3 and 0. 864 2, and the RMSECV and RMSEP were 0. 1600and 0. 2020mg g-1, respectively. The results show that compared with conventional iPLS, the ACO-iPLS can effectively select wavelength regions of near infrared spectroscopy and improve accuracy and robustness.

关 键 词：花茶 花青素 近红外光谱法 蚁群算法 区间偏最小二乘法 

分 类 号：O657.33[理学—分析化学]