我国牛肉大理石花纹等级图谱的改进研究

本文通过对牛肉眼肌截面照片的图像处理，提取其中大理石花纹的特征参数，建立了大理石花纹得分与大理石花纹总面积、大颗粒脂肪密度、中等颗粒脂肪密度和脂肪分布均匀度的线性预测模型，合理性较好（R2=0.8415）；并在此模型和有代表性的样本参数统计的基础上对不同大理石花纹等级的特征参数进行了量化，对我国《牛肉质量等级》中的大理石花纹等级图谱进行了调整，将原有的4级调整为7级，大理石花纹总面积百分比从1至7级依次为14%、11%、8%、6%、4%、2%和0.5%。

Improvement of Marbling Grading System for Chinese Beef

Features representing count and area densities of marbling were extracted from the marbling images. The linear model including sum marbling area, large marbling flecks density, middle marbling flecks density, and marbling distribution was made to predict marbling grade (R2=0.8415). Based on the prediction model and statistical data of features of marbling image, marbling grade was adjusted from four-point to seven-point rating scales. Sum marbling area from 1 to 7 point was 14%, 11%, 8%, 6%, 4%, 2%, 0.5% of the ribeye area, respectively.