System and method of predicting animal carcass characteristics using hyperspectral imaging

Abstract

Disclosed is a method for non-invasively predicting phenotype of an animal carcass that includes capturing one or more vertical or horizontal hyperspectral images of at least one of: an external or internal face of an entire carcass side of an animal through a hyperspectral imaging instrument; selecting a Region of Interest (ROI) representative of at least one of: the internal and external faces of the entire carcass side, and the entire carcass side of the animal; classifying each pixel for assigning corresponding pixel to a predefined constituent based on corresponding spectral information; extracting spectral information from each ROI by obtaining an average spectrum representative of an average chemical composition of the entire carcass of the animal; and developing one or more prediction or classification models for predicting one or more characteristics of the animal using one or more predictors, wherein the one or more predictors include at one of: output of the image segmentation, and the average spectra of the ROIs and optionally carcass hot weight.

Images