【摘要】 柑橘树锰、锌等缺乏问题普遍存在,但其症状有时不易识别,不利于对这些微量元素丰缺与否的诊断和有针对性的矫治.采用高光谱成像技术,研究了哈姆林甜橙叶片不同程度缺锰、缺锌症状的光谱响应特征及差异,对Fisher线性判别分析和最小二乘支持向量机两种识别模型的识别精度进行了比对分析.结果表明,利用高光谱原始全光谱反射率,结合最小二乘支持向量机建立叶片缺锰、缺锌症状识别模型,其建模集识别精度达91.88%,预测集识别率可以达到90.00%;利用连续投影算法筛选的40个特征波长,建立最小二乘支持向量机判别模型,其建模集识别精度为90.00%,预测集识别精度也达82.50%.表明基于高光谱成像技术对柑橘树缺锰、缺锌的准确识别是可行的,为高光谱成像技术应用于柑橘树缺素症的快速无损识别奠定了基础.更多还原

【Abstract】 In citrus orchards,manganese(Mn)and/or zinc(Zn)deficiencies are very common,while the symptoms caused by them are difficult to identify,which usually results in a wrong treatment during orchard management.In order to lay a foundation for application of hyperspectral imaging technology to the identification of nutrition deficiency in citrus,an experiment was made,in which the spectral response characteristics of‘Hamlin’Sweet Orange leaves lacking Mn and/or Zn in different extents were characterized with hyperspectral imaging technology,and then two recognition models were built using Fisher linear discriminant analysis(LDA)and least squares support vector machine(LSSVM),the recognition accuracy of which was subsequently compared.Our results showed that the model built by LSSVM with the whole spectral range wavelength gave a recognition accuracy of 91.88%in the modeling set and 90.00%in the prediction set;in addition,the model built by LSSVM with forty wavelengths selected by successive projections algorithm(SPA)resulted in a recognition accuracy of 82.50%.The results in this study indicated that it was feasible to identify Mn and/or Zn deficiency using hyperspectral imaging technology.更多还原