花生病害叶的傅里叶变换红外光谱研究

Study on Diseased Peanut Leaves by Fourier Transform Infrared Spectroscopy

[目的]探究傅里叶变换红外光谱(FTIR)技术在花生病害鉴别中的应用。[方法]利用FTIR技术并结合主成分分析和聚类分析方法对花生褐斑病叶、黑斑病叶、网斑病叶和正常叶进行鉴别。[结果]样品的红外光谱图整体相似,只在1 750～800 cm-1范围内吸收峰的位置和强度有微小差异。对原始光谱图做二阶导数处理,发现3 600～2 800 cm-1和1 750～650 cm-1范围内的二阶导数谱差异明显。选取3 600～2 800 cm-1和1 750～650 cm-1范围做主成分分析和聚类分析,所提取的3个主成分的累计贡献率达到94.9%,分类正确率达到100%,聚类分析的正确率达到94.6%。[结论]傅里叶变换红外光谱技术有望发展成为鉴别农作物病害的有力手段。

[ Objective] The paper was to study the application of Fourier transform infrared spectroscopy （ FrlR technology） in identification of peanut diseases. [ Method ] By using F＇FIR technology, combined with the methods of principal component analysis and hierarchical cluster analysis, the healthy leaves and three kinds of diseased leaves （cercospora black spot, cercospora brown spot and web blotch） were identified in the paper. [ Result] IR spectra of both diseased and healthy samples were similar, but tiny differences in wave-numbers and absorption in- tensities of peaks were observed in the range of 1 750 -800 cm-1. Significant differences were found in the second derivative spectra in the range of 3 600 -2 800 and 1 750 -650 cm- 1 which were selected to perform principle component and hierarchical cluster analysis. Three prin- cipal components had the cumulative contribution rate of 94.9% and yield classification accuracies of 100%. Hierarchical cluster analysis pro- vided about 94.6% of correct classification. [ Conclusion] Fourier transform infrared spectroscopy has a potential to be developed as a power- ful means for identifying crop diseases.