Pork grading is mandatory in meat production centres. It requires measurement of fat and lean layers thickness in the rib area of a pig carcass. The procedure is performed by manual insertion of an infrared probe into the carcass. It is desired to replace the probe insertion with ultrasound scanning to perform grading non-invasively, and to automate the procedure. A challenge is faced in positioning the ultrasound sensor normal to the skin at the scanning location.

In this thesis means of finding the surface normal on a pig carcass are developed. An ultrasound range sensor was found most suitable. Experimental studies conducted on the range sensor led to a proposed solution of using a single range sensor to perform skin surface following and detection of the normal to the surface. An algorithm was developed to achieve this goal. This thesis presents this algorithm along with simulation and experimental results.