Correlation of the Attributes Measured by Computer Vision with Moisture and Fat Content of Meat Batters

Abstract

The aim of this study was to determine if there is any correlation between moisture and fat content and such attributes estimated by the computer vision system (CVS) as white and red areas (%), values of colour coordinates in RGB and CIELAB colour systems, in the batters composed of porcine meat and fat (Experiment 1) or meat, fat and water (Experiment 2). The fat content (the Soxhlet method) was most highly correlated with the white fields’ area (r = 0.98, 0.85 and 0.85 for data obtained from Exp. 1, Exp. 2, and both, respectively). Also, the moisture content (the oven drying method) showed the strongest correlation with the area of white fields (r = −0.97, −0.85, −0.83, for Exp. 1, Exp. 2, and both, respectively). Thus, for estimation of fat and moisture content in meat batters the most useful CVS attribute is area of white fields.