Regresar

Soil NPK Levels Characterization Using Near Infrared and Artificial Neural Network

Abstract:

Soil, being the medium for plant growth serves as the primary nutrient source. Nitrogen (N), Phosphorus (P), and Potassium (K) are the vital nutrients plants need in the largest amounts. There should be sufficient quantities of these nutrients in the soil to ensure optimum plant growth resulting in high quality crops and improved crop yields. Farm managers then have a critical responsibility of assessing the NPK levels in soil. Soil chemical analysis was the reference method used to define soil NPK levels. Near Infrared (NIR) Spectroscopy was adopted as a non-destructive, fast, and environmentally friendly method for determining characteristic absorbance spectra of soil. The resulting variations in the measured NIR absorbance wavelengths ranging from 1240 nm to 1480 nm were evaluated and used to characterize the NPK nutrient levels of the various soil samples. Furthermore, the dataset was modelled and analysed using Artificial Neural Network (ANN) to determine the relationship of the NIR absorbance data to the soil NPK nutrient levels. Based on the regression analysis, the model's training R is 0.998, the testing R is 0.996, the test R is 0.996 and the overall R is 0.998. This study proves that NPK soil nutrient levels can be characterized in terms of NIR absorbance spectra.