Abstract: Two hundred and sixty-six brown rice samples were scanned with a FOSS TECATOR infratec 1241 grain analyzer and their grain quality characteristics were also determined by traditional methods.Using WINISI software,the models for predicting several grain quality characteristics of brown rice by near infrared transmission spectroscopy(NITS) were developed.The results indicated that the model for protein content prediction was the best,with the corrected standard error of performance SEP(C) being 0.312 and average determination coefficient of validation(RSQ) being 0.956;reasonable accurate models were also obtained for predicting the amylose content,alkali spreading value,transparency and chalky degree.Their SEP(C) were 1.672,0.389,0.034,4.024 and RSQ were 0.745,0.838,0.797,0.714,respectively.The repeatability of the NITS prediction and the application of the developed models in rice breeding program were also discussed.

Key words: near infrared transmission spectroscopy, brown rice, grain quality, model, methodology

摘要： 利用FOSS TECATOR 公司生产的1241型近红外透射光谱仪，对266份整粒糙米样品进行了光谱扫描并测定了其多项品质指标。借助于近红外定标软件WINISI，建立了利用近红外透射光谱同时测定糙米多项品质指标的模型。结果表明，蛋白质模型的分析效果最好，其外部检验 (经校正的) 工作标准误\[SEP (C) \] 为0.312，检验决定系数(RSQ )为0956，而其他4项指标直链淀粉含量、糊化温度(碱消值)、透明度（透光率）和垩白度模型的SEP (C)分别为1.672、 0.389、 0034和4.024，RSQ分别为0.745、0.838、0.797和0.714。分析了近红外测定的重复性并讨论了该模型在水稻遗传育种等方面的应用前景。

关键词: 近红外透射光谱技术, 糙米, 稻米品质, 模型, 测定方法