Abstract:

【Objective & Method】 Field experiments were conducted to assess the potassium content in leaves of Citrus sinensis (L)cv. Peng’an 100 Jincheng orange by using VIS/NIRS spectral method. Before calibration, principal component analysis (PCA) and partial least square (PLS) techniques were applied in data pre-processing. 【Result】 The order of leaf reflective spectrum intensity was K3＞K0＞K1＞K2＞K4 in the visible near-infrared range of 400-1000 nm, where the K fertilizer usage of K0, K1, K2, K3, K4 treatments were 0 g, 30 g, 75 g, 90 g, 120 g(k2O/plant/year) , respectively. The calibration models of potassium content were built by applying PLS and internal cross-validation test method and through processing the reflectance spectrum, the first derivatives, the second derivatives and the reciprocal logarithm spectrum of Peng’an 100 Jincheng leaves using multiplicative scatter correction(MSC). The results showed that the model of the second derivatives calibration of reflectance spectrum had the best predicative ability, the highest correlation coefficient, the smallest root mean square error of predictation (RMSEP) and the smallest absolute bias at 0.82, 0.0038 and -2.34E-05, respectively. 【Conclusion】 The second derivatives of reflectance spectrum could be used to predict the potassium content in Peng’an 100 Jincheng leaves. And 477-515 nm, 541-588 nm, 632-669 nm, 701-718 nm and 754-794 nm were the characteristics of wavelengths of second derivatives of reflectance spectrum predicting potassium content in summer shoot leaves of of Peng’an 100 Jincheng.

Key words: Jincheng orange leaf, potassium content, second derivatives, partial least squares (PLS), Vis/NIR-spectroscopy