Abstract:
IoT soil moisture sensors, especially capacitive types, play a vital role in precision irrigation due to offering cost-effective and rapid soil moisture monitoring. However, low-cost capacitive sensors often require accurate calibration specific to agricultural soil textures. This study focuses on calibrating Soil Stick sensors, commercially available in Thailand, using fine and medium-textured soils from agricultural areas in Phetchaburi Province. The calibration process establishes an equation relating the sensor's output voltage, connected to the NodeMCU ESP32 microcontroller board, to volumetric water content. The calibration results revealed a third-degree polynomial equation with an RMSE value of 0.07 cm3.cm-3. The performance evaluation was conducted using the calibrated Soil Stick sensor compared to the SM100 sensor, a factory-calibrated for various soil types. The Soil Stick sensor exhibited a lower RMSE value of 0.07 cm3.cm-3, whereas the SM100 sensor had an RMSE value of 0.08 cm3 cm-3. Furthermore, the confidence index of measurement (CI) demonstrated that the Soil Stick sensor achieved a value of 0.78, indicating a very good measurement performance, while the SM100 sensor yielded a CI of 0.66, denoting good measurement performance. The sensor's accuracy in soil moisture measurement was enhanced through calibration, enabling efficient control in irrigation applications.