在快速发展的中国,有机食品、绿色食品和无公害食品在产地环境、生产规程、加工、包装,尤其是生产资料的使用等方面都有严格区分,导致检验和认证需要花费大量的人力、物力、财力和时间.因此我们需要一种方便、精确的检测技术.基于化学计量学,本研究首次利用可见-近红外反射光谱快速检测有机胡萝卜、绿色胡萝卜和无公害胡萝卜.选择最佳区间(650-700 nm),三种分类模型包括主成分分析结合k均值聚类、主成分分析结合线性判别分析、反向传播神经网络鉴别126个胡萝卜样本的准确率分别为93.65%、100%,和98.41%.结果表明,可见-近红外光谱结合化学计量学技术可以有效地验证有机胡萝卜、绿色胡萝卜和无污染胡萝卜.为质量检验机构提供了一个可行的检测方法,从而促进包括有机食品、绿色食品、无公害食品等安全食品的发展.

Fast developing organic foods,green foods and non-pollution foods in China were strictly distinguished in the environment of producing area,production regulation,processing,packing,and especially the means of production.As a result,the inspection and certification cost a great deal of manpower,materials,financial resources and time.Therefore a convenient and accurate technique was needed for detection.In this work,visible and near infrared(Vis-NIR)reflectance spectroscopy,a rapid-online tool,was firstly used for detection of organic carrots,green carrots and non-pollution carrots based on chemometrics.Three classification models including principal component analysis combined with k-means clustering(PCA-KMC),PCA combined with linear discriminant analysis(PCA-LDA),and back propagation neural network(BPNN),built on the optimal region(650-700 nm),could identify 126 carrot samples with correct classification rates of 93.65%,100%,and 98.41%,respectively.The results indicated that Vis-NIR spectroscopy and chemometric techniques could efficiently verify organic carrots,green carrots and non-pollution carrots.It provides an available detection method for quality inspection agencies so as to promote the development of safe foods including organic food,green food and non-pollution foods.