摘要 |
PROBLEM TO BE SOLVED: To obtain a method for discriminating the quantity of edible meat according to a computer image analysis for discriminating the quality of meat accurately by improving the accuracy of binarization processing. SOLUTION: The image of the section of edible meat is divided into small regions that overlap one another with a maximum of 50%, the threshold of each small region is obtained by the discrimination binarization method according to a concentration histogram and is set to the threshold of the center point of each small region, the threshold of all points is determined according to the linear weighted sum that is inversely proportional to distance based on the threshold, all points (pixels) are binarized based on the threshold, the binary image of the entire section of the edible meat is obtained, a shape feature value paying attention to a fat cross-fertilization particle with a complex shape and that with a small area from the binary image, and the quality of the edible meat is discriminated according to the above information.
|