machine vision approach coupled with a hybrid ehd-convective dryer to model khalal slices drying process with anfis

Khalal is a product of date palm fruit before full ripeness and has a higher moisture content than rutab and fully ripened date fruit. this study deals with monitoring the real-time drying process of khalal thin slices in a hybrid electro-hydrodynamic (ehd)-convective hot air dryer. the real-time moisture ratio (mr) of khalal slices was estimated with an intelligent online machine vision system and eliminating the conventional weighing system was investigated. for this purpose, the samples were photographed at specified time intervals during the drying process. an adaptive neuro-fuzzy inference system (anfis) was developed to extract real-time models for the drying process. the input features contained different combinations of the temperature of the chamber, air velocity, and drying time along with the l*, a*, and b* coefficients of the image were calculated at different times. the performance of the developed models was evaluated, and the best model was selected. the results revealed that the differential sigmoid membership function with six inputs can provide the best estimation for the moisture ratio (mr) of the product with the coefficient of determination of 0.988 and 0.987 for train and test data, respectively. finally, it is concluded that the proposed online model can eliminate the need for an embedded weighing system through intelligent control of the ehd-convective dryer and provide a robust real-time prediction of the mr of khalal thin slices.