Application of visible and near infrared reflectance spectroscopy to predict total nitrogen in soil

Abstract

Aim : Methodology :Results : Interpretation :The application and use of precision agriculture technologies targeting reduction in agricultural inputs are associated with various raw data and their assessment to create application maps. Yield mapping and soil mapping are the two basic agricultural application maps for variable-rate applications. Creation of highresolution digital soil maps requires high resolution information. The objective of this study was to predict total nitrogen in soil, using visible and near infrared (vis-NIR) reflectance spectroscopy, which is a new technology used for making fast and accurate measurements of physical and chemical properties of soil.A set of 140 surface soil (0-20 cm) samples was collected from cultivated land, fruit, and vegetable fields distributed over Bursa province (Turkey). The diffuse reflectance spectra of soil samples were examined for total nitrogen content. Partial least squares regression (PLSR) analysis with full crossvalidation was performed to develop a model for total nitrogen. 2The prediction performance of the model produced 0.793 regression coefficient (R ),0.0274% root mean square errors of prediction (RMSEP) and 2.0802 residual prediction deviations (RPD).The results showed that the vis-NIR spectroscopy can be used to evaluate soil total nitrogen content. More research should be conducted to improve the model prediction of total nitrogen in soils with both high and low nitrogen content using vis-NIR reflectance spectroscopy.

Combining Variable Selection and Multiple Linear Regression for Soil Organic Matter and Total Nitrogen Estimation by DRIFT-MIR Spectroscopy

“…As effective alternatives to traditional chemical analysis, visible and near-infrared spectroscopy (Vis-NIR), mid-infrared spectroscopy (MIR), and combined diffuse reflectance spectroscopy have the potential to predict various soil properties simultaneously [5][6][7][8]. Diffuse reflectance spectroscopy in the Vis-NIR spectral range has been used widely to characterize SOM [9][10][11][12][13][14] and TN [13][14][15][16][17][18][19]. MIR has been demonstrated to predict SOM and TN, often with better accuracy than Vis-NIR-derived models [20][21][22][23][24][25][26][27].…”

Visible and Near-Infrared Spectroscopy Combined With Bayes Classifier Based on Wavelength Model Optimization Applied to Wine Multibrand Identification

“…This method usually does not require reagents, and it can measure samples directly, with the advantages of being quick and easy. Combined with the visible light region, visible-near infrared (Vis-NIR) spectroscopy has been applied in many fields, such as agriculture and food (1)(2)(3)(4)(5)(6), environment (7,8), and biomedicine (9)(10)(11)(12)(13)(14).…”