Soil water content estimated by resistive sensors - DOI: 10.7127/rbai.v1801367

Resumo

Agricultural automation is an alternative for more precise and sustainable agriculture. Sensors are used to collect information and contribute to monitoring and decision-making in production environments, mainly in irrigated agriculture. In this context, the objective of this study was to compare different models of soil moisture sensors under the same reading conditions. Two sensor models were evaluated: FC-28 (Sensor A) and HL-69 (Sensor B). The Student's t-test was applied at 5% significance level to compare the means of the readings between the sensor models (A and B). A study comparing sensors of the same model was conducted in a split-plot experimental design, in a 4×9 factorial arrangement (four sensors × nine readings) in sampling units (stainless-steel rings) with different soil water contents, with 95 replications. Linear response models with a plateau were used when the sensors showed reading variations as a function of the increase in soil water content. The data obtained were analyzed in the software R. The results showed that both sensors require independent, and a mathematical model should be developed for each sensor. The divergent results between the evaluated sensors denote low rigor in device standardization during manufacturing.