Real-time ultrasound image analysis for the estimation of carcass yield and pork quality

Abstract

Average backfat thickness measurements (liveweight of 92.5 kg) were made on 276 pigs using the Krautkramer USK7 ultrasonic machine. Immediately preceding and 1 h after slaughter real-time ultrasonic images were made between the 3rd and 4th last ribs with the Tokyo Keiki LS-1000 (n = 149) and/or CS-3000 (n = 240) machines. Image analysis software was used to measure fat thickness (FT), muscle depth (MD) and area (MA) as well as scoring the number of objects, object area and percentage object area of the loin to be used for predicting meat quality. Carcasses were also graded by the Hennessy Grading Probe (HGP). Prediction equations for lean in the primal cuts based on USK7 and LS-1000 animal fat measurements had R²-values (residual standard deviations, RSD) of 0.62 (27.0) and 0.66 (25.7). Adding MD or MA to LS-1000 FT measurements increased the R²-values to 0.68 and 0.66. Prediction equations using animal fat measurements made by the USK7 and CS-3000 had R²-values (RSD) of 0.66 (26.5) and 0.76 (22.4). Adding MD or MA to CS-3000 FT measurements made no further improvement in the R²-values. Estimation of commercial lean yield from carcass FT and MD measurements made by the HGP and LS-1000 had R²-values (RSD) of 0.58 (1.72) and 0.65 (1.56). Adding MA to LS-1000 measurements made no further improvement in the R²-values. Prediction equations based on carcass FT and MD measurements made by the HGPandCS-3000 had R²-values (RSD) of 0.68 (1.65) and 0.72 (1.54). Adding MA to CS-3000 measurements made no further improvement in the prediction equation. It was concluded that RTU has most value for predicting carcass lean content and further improvements in precision will come from more accurate FT measurements from RTU images made by image analysis software. Correlation of the number of objects, object area and of percent object area of image from RTU images with intramuscular fat or marbling score made on the live pig or carcass were low, and presently do not appear suitable for predicting intramuscular fat. Key words: Carcass composition, meat quality, marbling, intramuscular fat, sex, pigs