论文部分内容阅读
采用傅里叶变换红外光声光谱技术对10个品种的油菜籽样本进行品种鉴别。原始光声光谱卷积平滑后,首先采用全谱数据建立支持向量机鉴别模型,当RBF核函数的核参数γ值为0.01时,模型最大预测率为70%。利用方差分析的方法对全谱进行有效波长筛选,筛选后的波长用于建立支持向量机鉴别模型,当γ值取0.1时,模型的识别率和预测率均可达到100%。同时,采用偏最小二乘判别分析建立鉴别模型,作为支持向量机模型的对照,该模型的预测率仅为60%,明显低于支持向量机模型的预测精度。研究表明,红外光声光谱技术结合支持向量机,在油菜籽品种鉴别中有良好的应用性能。
Fourier transform infrared photoacoustic spectroscopy was used to identify 10 varieties of rapeseed samples. After the original photoacoustic spectrum was convoluted and smoothed, a support vector machine (SVM) discriminant model was first established based on the full spectral data. When the nuclear parameter γ of the RBF kernel is 0.01, the maximum prediction rate of the model is 70%. The variance analysis was used to select the effective wavelength of the whole spectrum. The filtered wavelength was used to establish the support vector machine (SVM) discriminant model. When the γ value was 0.1, the recognition rate and the prediction rate of the model could all reach 100%. At the same time, partial least square discriminant analysis was used to establish the discriminant model, which was compared with the support vector machine model. The forecasting rate of this model was only 60%, which was significantly lower than that of the support vector machine model. The research shows that infrared photoacoustic spectroscopy combined with support vector machine has good application performance in the identification of rapeseed varieties.