边缘检测、二值化处理进行牛肉分级的应用

Application of Edge Detection and Binarization for Beef Grading

采用图像处理技术自动估算牛肉眼肌横切面特征值,为基于计算机视觉的牛肉品质自动分级检测奠定基础。以牛胴体6～7肋横断面图像为试验材料,采用边缘检测、二值化处理技术等,运用VisualC++6.0编程语言,对牛肉眼肌的眼肌面积、脂肪、肌肉总面积比、脂肪分布均匀度、眼肌圆度、肌肉和脂肪色度值5个特征参数进行特征提取和检测。结果表明:经测量所得的眼肌面积越大,圆度越大,肌肉和脂肪色度值越高、大理石纹密度分布均匀的牛肉品质越好,相反,眼肌面积小、圆度小、肌肉和脂肪色度值越低、密度分布不均匀的牛肉品质低。该设计可有效计算眼肌面积和特征参数,能代替常规分级方法,实现牛肉质量等级的自动判别。

Automatic eigenvalue estimation of beef ribeye cross-section images through image processing lays the foundation for automatic beef quality grading based on computer vision technique. Digital images of the carcass cross section between the sixth and seventh ribs were subjected to feature extraction and detection of characteristic parameters （ribeye area, fat area ratio, total muscle area ratio, average fat distribution, the roundness of ribeye area and ribeye muscle and fat colors） by edge detection and binarization based on Visual C ＋＋ 6.0. Our results showed that larger ribeye areas had better roundness, higher chromatic values of muscle and fat and more uniform distribution of marbling, indicating better quality. On the contrary, lower-quality ribeyes had smaller areas, lower roundness values and chromatic values of muscle and fat and uneven distribution of marbling. The described design enables effective calculation of ribeye areas and characteristic parameters and consequent automatic identification of beef quality grades and can be an alternative to routine ~radin~ methods.