Development of an IoT-Based Soil Macronutrient Analysis System Utilizing Electrochemical Sensors and Machine Learning Algorithms

Efficient administration of principal soil nutrients is essential for achieving rational agriculture and minimizing economic losses. To improve on-site observing techniques and soil management, the utilization of advanced sensors capable of quantifying soil properties is crucial. In this study, three sensors were employed: an Ion Sensitive Field Effect Transistor (ISFET) pH sensor, a soil moisture sensor, and an RGB color sensor with an IR filter and white LED (TCS34725). The ISFET pH sensor, functioning as a Metal Oxide Semiconductor Field Effect Transistor (MOSFET), provided pH measurements ranging from 1 to 14 and exhibited temperature resistance between 0 to 100 degrees Celsius. The soil moisture sensor was designed to determine soil moisture content within a range of 0 to 1023. The RGB color sensor, by reducing the infrared spectral component of incoming light, enabled precise color measurements. These sensors were integrated into the Arduino MEGA 2560 microcontroller board to facilitate simultaneous operation of complex schemes. The acquired sensor data were correlated to the concentrations of soil macronutrients, specifically Nitrogen (N), Phosphorus (P), and Potassium (K). Through the implementation of machine learning algorithms, the levels of macronutrients can be determined, and for the training set of these algorithms, data collection was conducted using over 300 soil samples.