【摘要】 为研究近红外光谱分析技术结合支持向量机(SVM)模式识别原理在鉴别烟叶产地中的可行性,选择云南省曲靖市所辖6个不同区县的180个烟叶样品为研究对象。对其近红外光谱进行一阶求导后,通过matlab分析软件建立预测各烟叶产地的近红外模型,并应用该模型对78个烟叶样品进行实际产地预测。实验结果表明:上述方法建立的预测模型训练样本的正确识别率100.00%,测试样本正确识别率为91.03%。结果显示,运用基于SVM的近红外光谱技术可以实现烤烟烟叶产地的快速准确判定。更多还原

【Abstract】 In order to study its feasibility of determining the producing area of tobacco leaf using near-infrared reflectance spectroscopy with support vector machines(SVM),180 tobacco leaf samples which came from 6 different county of Qujing,Yunnan were chosen as the target.The prediction model was built by matlab software after first derivation of near-infrared spectroscopy,and then the model was applied to predict the producing area of 78 samples.The experimental result showed that the classification accuracies for training set and test set of the SVM model were 100% and 91.03%,showing that near-infrared with SVM is a fast and precise method to discriminate the producing area of tobacco leaf.更多还原