Title :
Modelling primal cuts on pork carcasses for automated butchery
Author :
Wadie, I.H.C. ; Khodabandehloo, K.
Author_Institution :
AMARC, Bristol Univ., UK
Abstract :
To separate half carcasses into primals automatically the required cut paths need to be determined. The neck meat is removed first, then a series of skin cuts are performed which follow the bone cuts on the medial plane of the carcass. The depth of these cuts are being determined by a surface model. The loin skin and shoulder are then removed without tearing any meat from the loin primal. The shoulder is removed by cutting behind the scapula once the loin skin has been cut away. The bone cuts are then performed to separate the remaining primals. These cuts are performed currently by trained butchers to achieve primals which are acceptable to the consumer and which maximise the profit from each carcass. In order to automate them, they have been modelled from anatomical features on the carcass. Some cuts pass directly through anatomical points (anatomical cuts) of which the 4th/5th bone and ham bone cuts have been modelled simply whilst other cuts do not pass through anatomical points and are modelled by relating them to relevant anatomical features (nonanatomical cuts with anatomical descriptions), an example of this type of cut is presented in this paper. The alternative form of modelling is to generate general cut paths which can be scaled according to general carcass dimensions but the use of anatomical points is preferred because the models adapt naturally due to the difference in size between each carcass and the natural variation in the position of the features on each carcass. Robots with vision are used
Keywords :
cutting; food processing industry; industrial robots; process control; robot vision; anatomical features; automated butchery; computer vision; meat; pig carcasses; pork; primal cuts; robot;
Conference_Titel :
Intelligent Automation for Processing Non-Rigid Products, IEE Colloquium on
Conference_Location :
London