Intelligent automation of bird deboning

Second and further processing operations constitute the largest use of on-the-line labor in today's poultry processing plants. Deboning is a significant part of these operations. This paper describes continuing work in the intelligent automation of bird deboning, a task that is conceptualized into three parts: characterizing the non-uniform bird product (using statistical studies and image processing), obtaining a nominal cutting path (using image features correlated to internal bird structures and robot kinematics), and correcting for deviations from the nominal cutting path (using force control). Results from deboning experiments using a prototype system developed to perform bird shoulder cuts are reported.