PREDICTION OF TEXTURE CHARACTERISTICS IN APPLE DRYING USING COMPUTER VISION

Abstract

It was proved that computer vision is applicable for continuous estimation of bulk volume from diameter imaging of apple slices and an empirical model with an average relative percent error of approximately 6.45 % was established, which allowed continuous estimation of total porosity as a function of volume shrinkage and moisture content. Texture parameters were dependent on porosity evolution and glass transition, which was a function of drying temperature and moisture content. Glass transition occurred within moisture content range of 1.0 g/g to 0.26 g/g when open-pores converted into closed-pores. To create crispy texture, the drying process should be carried out at 80 ºC to moisture content of less than 0.5 g/g with a total porosity of more than 0.7 when materials undergo glass transition and case hardening. This knowledge makes it possible to develop an optimal industrially-scaled drying process, using computer vision for continuous monitoring and prediction of texture.