Quantitative determination and characteristic wavelength selection of available nitrogen in coco-peat by NIR spectroscopy

Abstract

Available nitrogen content reflects the nitrogen latest supply capacity of coco-peat. Rapid determination of available nitrogen content in coco-peat is of great significance for precision fertilization in precision agriculture. In this study, available nitrogen of coco-peat was rapidly and quantitatively detected by near-infrared (NIR) spectroscopy. The mathematical model between spectral data and measured available nitrogen values was established, and the impacts of different spectral pre-processing on the whole band modeling results were compared and analyzed. Successive projections algorithm (SPA) and elimination of uninformative variables (UVE) were applied to remove redundant and irrelevant information variables from original spectrum and various pre-processing spectra, and the corresponding PLSR and MLR spectral prediction models were established. Comparing the modeling results of using full-band spectral data and eliminating the irrelevant information spectral data, the optimal spectral prediction model was established by using MLR. And the model with filtered characteristic wavelength spectral data by SPA after Savitzky-Golay smoothing (SG smoothing) pre-processing yielded optimum results with the correlation coefficients for calibration set (RC) 0.996 and prediction set (RP) 0.990; the root mean square errors for calibration set (RMSEC) and prediction set (RMSEP) were 4.634 mg/100 g and 7.203 mg/100 g, respectively; and the ratio of prediction to deviation (RPD) was 7.011. The results showed that fast quantitative detection of available nitrogen of coco-peat could be achieved by NIR spectroscopy, and it could provide a reference for the rapid detection of substrate fertilizer content in facility agriculture.