Predição da composição da carcaça e da qualidade da carne de cordeiros deslanados utilizando escores de tipificação do sistema europeu e video image analisys

Abstract

This study aims to propose a methodology of class determination based on Video Image Analysis (VIA) able to express the dimensions, profiles and distribution of hair lambs carcass fat cover, comparing it with the usual methodology of the European Classification and Tipification of Carcasses from the evaluation of their capacity to predict quantitative characteristics, regional composition and tissue composition. Data of 55 lamb’s carcasses of Santa Inês and Dorper x Santa Inês genotypes from commercial flocks in Pará State, finished in confinement at the Federal Institute of Education, Science and Technology of Pará Campus Castanhal, will be used. Carcasses were typified for conformation and fatness according to the European System. A set of images of cold carcasses obtained through digital camera was processed to get shape descriptors. Carcass weights, cold carcass weights, weights of the commercial cuts shoulder, leg, rib and loin; tissue, muscle, bone and fat components of the carcass and their relationships; and CIELab* staining, cooking loss and shear force of Longissimus lumborum muscle samples was obtained. A minimum set of shape descriptors representative of the size and distribution of carcass fat and prediction models of shape, profile and finish classes was obtained, a methodology called Digital Accessor Vision (DAVis), which was able to predict the main characteristics (carcass weights, shank, total muscle, edible portion), and most of the meat quality characteristics (a *, b *, C *, shear force). The result allows predictions with superior adjustments and lower errors than the European System, based on the visualization of a human acessor, being a promising technique in the classification of lamb carcasses by image processing.