番茄游离氨基酸总量近红外光谱预测模型研究  

Research on predictive models of total free amino acid in tomato by near infrared spectroscopy

作　　者：李天华[1,2] 陈龚飞 刘光伟[3,4] 魏珉[5] 侯加林[1,2] 

机构地区：[1]山东农业大学机械与电子工程学院,山东泰安271018 [2]山东省园艺机械与装备重点实验室,山东泰安271018 [3]山东省泰安市岱岳区产品质量监督检验所 [4]山东省泰安市产品质量监督检验所,山东泰安271000 [5]山东农业大学园艺科学与工程学院,山东泰安271018

出　　处：《计算机与应用化学》2014年第4期494-498,共5页Computers and Applied Chemistry

基　　金：山东省现代农业产业技术体系建设专项(SDARS-2010-2-3-1)

摘　　要：采用近红外光谱分析技术,建立番茄中游离氨基酸总量预测模型。从一至三穗成熟果实中共采集番茄108个,其中84个做校正集,24个做验证集。从数据归一化、数据格式、数据平滑几个方面选择不同的光谱预处理方法,确定最佳方法为:"Mean Centering"+"Second derivative"+"Norris derivative filter"。将全波数范围(40000~11000)cm^(-1)划分为70个区间,得到最佳建模区间组合为9,10,17,56,57,61。利用偏最小二乘法建立预测模型,得到相关评价指标R、RMSEC、RMSEP及模型预测准确率分别为0.936、6.72μg/100g、7.15μg/100g和92.5%。评价指标及对验证集的预测结果表明,所建模型用来实现对番茄中游离氨基酸总量进行无损、快速预测是可行的。Using near infrared spectral analysis technology, a predictive model was established about total free amino acid in tomato. 108 samples were acquired from one to three ear ripe fruits, 84 of which constituted the calibration set, 24 samples constituted the validation set. Different spectra pretreatment methods were chosen from the data normalization, data format and smoothing aspects, the optimal method was decided “Mean Centering” ＋“Second derivative” ＋“Norris derivative filter”. 70 sections were divided in （4000-11000） cm-1, get the best combination of intervals was 9, 10, 17, 56, 57, 61. Prediction model was set up by using PLS, the related evaluation index R, RMSEC, RMSEP and PI were 0.936, 6.72, 7.15 and 92.5 %. Evaluation index and the validation of prediction results show that the model is feasible, which is used to realize nondestructive and rapid prediction on the amount of free amino acid in the tomato.

关 键 词：番茄 游离氨基酸 近红外分析 预测模型 偏最小二乘法 联合区间法 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]