The use of computer vision system to detect pork defects

Abstract The aim of this study was to determine the effectiveness of computer vision system (CVS) to detect meat defects of m. longissimus lumborum ( LL ) in industrial settings. The material consisted of 230 muscles. Based on pH 1 (45 min) and pH 2 (24 h post-mortem) meat classification into quality groups was conducted. To give more precise characterization of the raw material (proving the defect or not) the electrical conductivity (EC), drip loss, thermal drip and water holding capacity (WHC) were determined. The color of the meat in CIEL*a*b* and by CVS was measured and the study into how the CVS can be employed in meat defect detection was done. It was found that it is possible to employ the CVS to detect PSE (pale, soft, exudative) and DFD (dark, firm, dry) and to classify meat into quality groups. It was not possible to differentiate RSE (red, soft, exudative) from RFN (red, firm, normal) meat in this study. The highest accuracy of raw material classification using the CVS method was reported for the HSL (hue, saturation, lightness) color parameters at 81.7%. Therefore, the computer vision system can be employed for rapid analysis of the quality of pork m. longissimus lumborum under industrial conditions.