【摘要】 目的:建立快速鉴别及评价薏苡仁及其伪品草珠子种仁的方法。方法:采用衰减全反射傅里叶变换红外光谱(ATR-FTIR)法测定薏苡仁与草珠子种仁的红外光谱,选择4个特征吸收峰所对应的吸光度值作为特征向量,建立人工神经网络(ANN)模型,对不同薏苡仁与草珠子种仁样本所得红外光谱的特征向量进行分类。结果:所建立的方法对薏苡仁与草珠子种仁的识别率均为100%。结论:该方法判别准确度高,可用于薏苡仁与草珠子种仁的快速鉴别。更多还原

【Abstract】 OBJECTIVE:To establish the method for rapid identification and evaluation of Coicis Semen and Coix lachrymal-jobi. METHODS:The infrared spectrum of Coicis Semen and C. lachrymal-jobi were identified by ATR-FTIR. Using absorbance value of 4 characteristic absorption peaks as feature vector,artificial neural network(ANN)model was established to classify feature vector of infrared spectrum of Coicis Semen and C. lachrymal-jobi. RESULTS:Established method was highly accurate,and the correct rates of Coicis Semen and C. lachrymal-jobi identification were both 100%. CONCLUSIONS:The method has high accuracy,and can be used for rapid identification of Coicis Semen and C. lachrymal-jobi.更多还原