Alert System to Prevent Damage During Drying of PGI Peppers in Southern Italy: Preliminary Results

The Pepper of Senise PGI is a typical product of Basilicata region in southern Italy. Historically, the drying process of ‘Peperone Crusco’ took place during the summer months, in the wind and under the porches of rural buildings. This method is not in line with today’s business needs (quantity, quality, traceability and healthiness of production) but it is necessary to follow the strict requirements of the production specification to maintain the PGI denomination. The objective is to develop a low-cost monitoring system suitable for the regulation that reduces product losses during the drying phase due to invasive rotting phenomena (losses exceeding 20% of dry matter). To monitor outdoor agrometeorological variables, the Senise ALSIA weather station, the closest to the greenhouse, was use, while, another equipped with air humidity and temperature sensors was positioned inside the drying greenhouse. A feedforward neural network (FFNN) was trained using climate parameters collected from the empty greenhouse and data from the Alsia outdoor weather station, employing a multivariate approach. The goal was to predict in advance humidity and temperature inside the greenhouse, as it was empty, also when the pepper was introduced in the greenhouse. To evaluate the impact of the presence of the peppers on environmental conditions, the predicted parameters were compared with the measured values recorded in the greenhouse after the peppers were introduced. From the comparison between the predicted and measured parameters it was possible to identify the time intervals in which the values of humidity and temperature became higher to intervene by improving the natural ventilation of the greenhouse. The system was trained and tested during 2022 year. The developed neural network can predict with a good accuracy level (R2 = 0.96) the microclimate variables inside the greenhouse during the drying period.