Sensor Development for Soil-Property Detection Using Near Infrared Spectroscopy

A utilization of Near-Infrared Spectroscopy (NIRs) for prediction of soil properties is a cost and time effective method. Spectral data are often suitable for estimation of biochemical soil quality indicators. This paper introduces a sensor development for soil-property detection using near infrared spectroscopy to avoid dangerous heavy metals and guarantee agricultural product quality. Such as arsenic in contaminated soils transporting to human body via vegetables or other products. In order to create a predictive model, experiments are conducted with samples that added exact quantity of arsenic trioxide (AS2O3). Furthermore, various soils from agricultural sites are also collected to perform experiments. Partial Least Square Regression (PLSR) is used as a signal processing method in order to create predictive models for common dangerous element in soils, such as iron and arsenic.