Nondestructive Determination of Leaf Nitrogen Content in Corn by Hyperspectral Imaging Using Spectral and Texture Fusion

The nitrogen content is an important indicator affecting corn plants’ growth status. Most of the standard hyperspectral imaging-based techniques for nondestructive detection crop use a single feature as input variable model, which reduces generalization ability prediction model. To this end, model leaves based on fusion image and spectral features proposed. In study, at modulation stage were studied, samples with different levels numbered, their data in wavelength range 400~1100 nm collected. average spectrum models was used valid information. First-order derivatives, normal variables transformation (SNV), Savitzky-Golay (S-G) smoothing, normalization selected to preprocess features. CARS-SPA algorithm screen sensitive variables. gray level co-currency matrix (GLCM) chosen extract texture test samples. Corn leaf fused modeled target Partial least squares regression (PLSR) support vector machine (SVR) predict leaves’ content. results showed that spectral-based improved performance some extent compared univariate models. PLSR predicted best results, RP2 RMSEP 0.987 0.047. This method provides reliable theoretical basis technical developing accurate leaves.