RMSE)為2.51,。It was studied that using spectral analysis to quantitatively analyze the rape total nitrogen content. Stepwise regression was used to select the characteristic wavelength of total nitrogen content against rape leaf spectra for nitrogen content prediction. The method of principal component analysis (PCA) was used to avoid the effect of multiple co-linearity among the spectral data. In order to enhance model forecast precision, the method of support vector machine regression(SVR)was used to establish the model between the rape total nitrogen content and the spectral characteristic wavelength data. From the rape spectral data under different nitrogen level, it was found that the linear relationships between rape total nitrogen content and spectral reflectance value of 406, 460, 556, 634, 662, 675nm are very notable. The correlation coefficient between the predict value and the real value is 0.89, the RMSE is 2.51.
參考文獻(xiàn)
相似文獻(xiàn)
引證文獻(xiàn)
引用本文
張曉東,毛罕平,程秀花.基于PCA_SVR的油菜氮素光譜特征定量分析模型[J].農(nóng)業(yè)機(jī)械學(xué)報(bào),2009,40(4):161-165. Content Spectral Character Models Based on PCA_SVR Method[J]. Transactions of the Chinese Society for Agricultural Machinery,2009,40(4):161-165.