Abstract: Beef marbling is the most important parameter of carcass grading and beef marbling grading is based on the measurement of fat density in the muscle of eye area. An automatic beef marbling grading technology was explored based on image processing. Imaging acquisition, single threshold segmentation, morphological erosion and expansion were used to extract feature parameters of images. Multiple regression analysis method was used to classify beef marbling structure and the final grading evaluation was implemented using an automatic grading software developed based on VC++. Our grading model had good accuracy and robustness.

Key words: beef, marbling, image processing, mathematical model