近红外透射光谱法(NITS)分析大豆品质的研究  

STUDY ON ANALYSIS OF SOYBEAN QUALITY BY NEAR INFRARED TRANSMITTANCE SPECTROSCOPY

作　　者：姚鑫淼[1] 张瑞英[1] 李霞辉[1] 程爱华[1] 张晓波[1] 高春霞[1] 

机构地区：[1]农业部谷物及制品质量监督检验测试中心,哈尔滨150086

出　　处：《大豆科学》2006年第4期417-420,424,共5页Soybean Science

基　　金：2005国家大豆良种推广补贴项目

摘　　要：为了研究大豆品质性状的快速测定方法,以我国东北部四省区572份大豆样品为材料,采用近红外透射(NITS)技术非破坏性测定大豆的粗蛋白质、粗脂肪含量。调选出的校正集样品经实验室常规分析测定,建立其吸收光谱与化学成分间的关系模型,校正并优化原有测定方程。校正方程经预测获得了较高的预测集决定系数0.9757(蛋白质)、0.9549(脂肪)和较低的标准误差2.18(蛋白质)、0.88(脂肪)。结果表明近红外透射光谱技术可广泛应用于大豆品种品质普查、大豆品质育种材料筛选和商品大豆质量分类分级。本研究试验用样品数量多、来源分布均匀、品种信息丰富,因此,所建立的近红外透射预测模型适用范围广。To study the rapid analysis method for soybean quality, 572 soybean variety samples were collected from four provinces in Northeast China. Soybean protein and oil content were analyzed by near-infrared transmittance spectroscopy without destroying. Calibration model were established with data from chemical analysis and absorbed spectrum of calibration samples to improve the accuracy of prediction results. Satisfied r^2 0.9757 （protein）, 0.9549 （oil） and SEP 2.18（protein）, 0.88（oii） were achieved. Results showed that near-infrared transmittance spectrum techniques could be applied in soybean quality determination, generation selection in soybean breeding program and soybean quality classification. Abundant samples and varieties applied in this research made the prediction model more practicable.

关 键 词：大豆 近红外透射 校正 预测 

分 类 号：S565.1[农业科学—作物学]