AVALIAÇÃO DOS CORTES COMERCIAIS E DAS CARACTERÍSTICAS DA COMPOSIÇÃO TECIDUAL DA CARCAÇA DE CORDEIROS DESLANADOS ATRAVÉS DO VIDEO IMAGE ANALYSIS

The objective of this study was to analyze if the Video Image Analysis (VIA) technique applied in dorsal images of sheep carcasses is capable through the use of filters, extract real information from the carcass and if through a modeling it is able to define relations between the images and the real characteristics. Quatitative parameters and images were obtained. The VL's: CARCASS, CUTS AND IMAGE; composed by VM's. For the reduction of VL data, the cross loading values were considered to meet the Average Variance Explained (AVE). The VIA associated with multivariate modeling established relationships between the VL's, however, pointed out a negative correlation between the VL IMAGE and the others because the filters used were inefficient in extracting information from muscle and fat. The efficiency of the VIA depends on the filter used was not able to extract characters from the image, but multivariate modeling structured the relation between the image and the real character.