A portable soil nitrogen detector based on NIRS

Abstract

As one of the most important soil nutrient components, soil total nitrogen (TN) content needs to be measured in precision agriculture. A portable soil TN detector based on the 89S52 microcontroller was developed, and a Back Propagation Neural Network (BP-NN) estimation model embedded in the detector was established using near-infrared reflectance spectroscopy with absorbance data at 1550, 1300, 1200, 1100, 1050, and 940 nm wavelengths. The detector consisted of two parts, an optical unit and a control unit. The optical unit included six near-infrared lamp-houses, a shared lamp-house drive circuit, a shared incidence and reflectance Y-type optical fiber, a probe, and a photoelectric sensor. The control unit included an amplifier circuit, a filter circuit, an analog-to-digital converter circuit, an LCD display, and a U-disk storage component. All six absorbance data as inputs were used to calculate soil TN content by means of the estimation model. Finally, the calculated soil TN content was displayed on the LCD display and at the same time stored in the U-disk. A calibration experiment was conducted. The soil TN content correlation coefficient (R 2) of the BP-NN estimation model was 0.88, and the validation R 2 was 0.75. This result indicated that the developed detector had a stable performance and a high precision.