Vol 62, No 2 May 2009 International Journal of Dairy Technology 147 ORIGINAL RESEARCH *Author for correspondence. E-mail: hilde.kraggerud@ tine.no © 2009 TINE BA. Journal compilation © Society of Dairy Technology doi: 10.1111/j.1471-0307.2009.00478.x Blackwell Publishing Ltd Oxford, UK IDT International Journal of Dairy Technology 1364-727X 1364-0307 Society of Dairy Technology 2009 XXX ORIGINAL RESEARCH ORIGINAL RESEARCH X-ray images for the control of eye formation in cheese H KRAGGERUD,* 1,2 J P WOLD, 3 M HØY 3 and R K ABRAHAMSEN 1 1 Department of Chemistry, Biotechnology and Food Science, Norwegian University of Life Sciences, N-1432 Ås, Norway, 2 TINE R & D Center, Box 50, N-4358 Kleppe, Norway, and 3 Nofima Mat, Norwegian Food Research Institute N-1432 Ås, Norway There is demand for non-destructive monitoring of eye formation in cheese during ripening. The objective of this work was to develop a simple method based on existing equipment in the dairy industry. Images were acquired using a conventional, low resolution online X-ray instrument. Image processing methods for detecting eyes of cheese and measuring volume and size distribution were developed. Sufficient detection of overlapping eyes was obtained. Semihard cheese with propionibacteria ripened under different conditions was analysed. The method was found promising for quality control as it will make possible non-destructive monitoring of eye formation of cheese throughout the ripening period. Keywords X-ray, Image processing, Eye formation, Cheese, Ripening temperature. *Author for correspondence. E-mail:hilde.kraggerud@tine.no INTRODUCTION The size of the eyes is an important quality parameter for large-eyed cheese varieties. Product specifications typically contain measures of the number and the size of the eyes. Cheese makers try to control this quality parameter by evaluating the cheese during warm room ripening, traditionally done by listening to the sound made while tapping the surface of the cheese, in combination with ‘looking inside’. This can be done by cutting the cheese in two halves, or by taking out a small cylinder from the cheese using a cheese trier. In both cases the cheese cannot be used for normal commercial sale after examination. A major disadvantage is also that only a small part of the cheese volume is examined. Fermentation of lactose by heterofermentative lactic acid bacteria in the starter culture is a source of CO 2 formation in the cheese during cheese- making and ripening. The quantity of gas formed in this type of fermentation is relatively small. In cheese varieties with propionic acid bacteria as an adjuct culture, the classic propionic acid fermentation was described as essential for the formation of CO 2 over 100 years ago, and thus essential for the develop- ment of characteristic eyes in such cheese varieties (von Freudenreich and Orla-Jensen 1906). However, the metabolism of propionibacteria in cheese is complicated and not yet fully understood. The two main metabolic pathways of CO 2 formation from lactate by propionibacteria are the classic propionic acid fermentation, in which lactate is converted to propionate, acetate and CO 2 , and amino acid catabolism, a pathway that was discovered later (Brendehaug and Langsrud 1985: Frölich-Wyder and Bachmann 2004). Supersaturation with CO 2 gas is needed for eye formation and can be achieved when the rate of CO 2 production is relatively fast and the rate of diffusion out of the cheese is slow. Part of the CO 2 gas will remain in the eyes of the cheese, some CO 2 will remain dissolved in the body of the cheese and some will diffuse out of the cheese loaf (van den Berg et al. 2004). When vacuum packaging in plastic films is used during ripening of rindless cheese, the diffusion rate of gas through the film will affect the gas pressure and thereby possibly influence eye formation. Gas formation by propionibacteria primarily takes place during a warm room stage of ripening. When the development of eyes is sufficient, pro- pionic acid formation is retarded by cooling the cheese to a lower temperature (Frölich-Wyder and Bachmann 2004). X-rays have been used for imaging for many years. The best known use of X-rays is within medical diagnostics but they are also used exten- sively in industry and other areas. The intensity of the X-rays is modified by absorption by the material they are passing through and the resulting energy is captured by a detector to form an image in grey scales (Gonzales and Woods 2001). Image processing techniques have been used increasingly for food quality evaluation in recent years. Image features like colour, size, shape and texture have been applied in various food- monitoring applications. Different methods of image acquisition, using digital cameras, ultrasound,