Computer vision for bulk volume estimation of apple slices during drying

The accuracy of imaging for bulk volume estimation of apple slices in the process of drying at temperatures from 40 to 80°C was investigated and compared with physical caliper measurements. The initial hypothesis was to estimate bulk volume of cylindrically shaped apple slices from diameter and thickness imaging. Imaging of diameter showed strong agreement with caliper measurements throughout the entire drying process, however imaging of thickness was not accurate due to the irregular shrinkage and bending phenomena. A linear model reflecting the relationship between diameter and bulk volume was developed and validated at temperatures of 40, 60, and 80°C. It showed good correlation between imaging diameter and bulk volume changes in the range of moisture contents from 9.1 to 0.25?g/g. The relative percentage error of the prediction model was approximately 6.45%, which suggests that imaging can be used as a robust tool for bulk volume estimation of apple slices during convective drying.