论文部分内容阅读
[目的]研究选取稻飞虱为害后水稻叶片的有效特征波段,用于从大量成像光谱数据中快速识别和分类稻飞虱为害后的水稻叶片。[方法]实验采用多光谱成像系统对400~720 nm波段范围,每隔5 nm的稻飞虱为害后的水稻叶片进行多光谱成像。[结果]根据波段指数原理,计算得出波段515,510,710,555,630,535,505,530和595 nm具有较理想的波段指数值,这些波段信息量丰富、相关性小;实验通过两种分类方法分别对稻飞虱为害后的水稻叶片的分类精度予以计算,得出全波段和特征波段的分类精度均大于90.00%。[结论]这些选取的波段可以作为稻飞虱为害后水稻叶片的有效特征波段,可以用于从大范围农作物中快速识别和分类水稻叶片。
[Objective] The research aimed to study the effective characteristic wave band of rice leaves after planthoppers infestation, which is used to quickly identify and classify the rice leaves damaged by planthopper from a large number of imaging spectral data. [Method] The multispectral imaging system was used to perform multispectral imaging of rice leaves damaged by planthoppers every 5 nm in the wavelength range of 400-720 nm. [Result] According to the band index principle, the calculated band indices 515, 510, 710, 555, 630, 535, 505, 530 and 595 nm had the ideal band index values. These bands were informative and the correlation was small. Two kinds of classification methods were used to detect the effects of rice planthopper The accuracy of the classification is calculated, and the classification accuracy of the full band and the characteristic band are all greater than 90.00%. [Conclusion] The selected bands could be used as effective bands of rice leaves damaged by planthoppers, which could be used to quickly identify and classify rice leaves from a wide range of crops.