Error Analysis in Detection of Soil Moisture Using Triad Spectroscopy Sensor

Soil moisture value is critical to the global hydrological cycle and impacts agriculture,environmental science, urban planning,forestry,and more. Traditional methods of measuring soil moisture, though well established, are time consuming, labor intensive, requires installation and maintenance of equipment, expensive and gives low resolution outputs. This calls for the requirement of a soil moisture calculation method which is instantaneous, portable,economic and efficient. Spectroscopy has emerged as a promising solution since it can deliver valuable insights into the composition, structure, and properties of materials across diverse fields. In this study, a spectroscopic sensor was employed to analyze soil samples,and its readings were compared with soil moisture values determined through the conventional gravimetric method. By applying different regression approaches, a relationship between the sensor output and soil moisture was established, aiming to minimize errors such as mean absolute error (MAE), mean bias error (MBE), and root mean square error (RMSE).The findings revealed that a third-order polynomial regression equation at 705 nm wavelength provided the most accurate results, with error values of MAE, MBE, and RMSE being 0.38, 0.06, and 0.5, respectively. This outcome demonstrates that combining the spectroscopic sensor with the derived equation can offer an instantaneous, non-contact method for soil moisture estimation. Compared to the lengthy and labor-intensive gravimetric process, this approach provides a rapid and practical solution for soil moisture assessment in various applications.