Two-dimensional modelling of pig carcass spines for robotic evisceration

State-of-the-art slaughter techniques are still mainly manual. A few attempts have recently been made to automate the primary pig slaughter process focussing on the partial processes before and after the evisceration process. This paper deals with two-dimensional modelling of the spine of pig carcasses as a means of control of robotic evisceration of pigs by predicting the path needed to cut the connective tissue which attaches the intestines to the spine. A generic model has been produced based on a representative sample of geometrical measurements of 29 carcasses including the carcass length and depth. It is estimated that 99.8% of the predicted cut paths are accurate to within 14.2 mm which is sufficient to guide a compliant tool along the spine to severe the connective tissue. A concept to improve the path prediction by grouping the carcasses was tried. However, errors in predicting the group of a carcass negated the improved accuracy achieved. The best form of prediction of the carcass spine was found to determine the parameters of a single six-parameter model from the length and depth of the carcass.