【摘要】 本文建立了近红外透射技术 (NIT)分析大豆样本的新方法 ,批量地分析了大豆子粒的水分、蛋白、脂肪 ,定标误差分别为 0 .36 9、0 .5 78、0 .4 6 0 ,相关系数分别为 0 .95 8、0 .913、0 .912。通过 30个样本的独立样品集的定标验证 ,预测误差分别为 0 .4 0 4、0 .6 2 2、0 .5 0 8,相关系数分别为 0 .95 2、0 .90 1、0 .90 8。结果表明 ,近红外光谱技术用于测试大豆品质是可行的 ,可用于育种的早代选择更多还原

【Abstract】 The paper describes a new method for the analysis of soybean kernels by near-infrared transmission(NIT) technique.The calibration error(SECV) for moisture,protein and fat in soybean whole kernels are 0.369,0.578 and 0.460,respectively,with correlation coefficient of 0.958,0.913 and 0.912,respectively.The method was tested for moisture,protein and fat in 30 samples with the prediction error(SEP) of 0.404,0.622,0.508 and correlation coefficient of 0.952,0.901,0.908,respectively.The result indicates NIT is feasible for fast analysis of whole soybean kernels and can be used in breeding research.更多还原