牛胴体眼肌切面分级特征的提取与分析  

Methods of extraction and analysis of grading features of ribeye section of beef carcass

作　　者：孟祥艳[1] 孙永海[1] 王笑丹[1] 李志[1] 胡铁军[2] 何兵[2] 

机构地区：[1]吉林大学生物与农业工程学院,长春130022 [2]吉林省长春皓月清真肉业股份有限公司,长春130062

出　　处：《吉林大学学报（工学版）》2010年第2期485-490,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：吉林省科技厅项目(20060217);吉林省人民政府人才开发基金项目

摘　　要：利用具有USB接口的CMOS摄像头和VFW技术构建了牛胴体分级信息自动采集系统。提出眼肌切面分级特征的测量和分析方法,采用计算机图像分析理论提取眼肌面积、背膘厚度、大理石花纹等级,并提出了评价大理石花纹等级的指标。背膘厚度的提取准确率平均值为92.72%,标准偏差为4.75;眼肌区域提取准确率的平均值为94.84%,标准偏差为2.52。利用BP神经网络对大理石花纹的等级进行建模、验证,准确率达到了94.74%,满足牛胴体分级信息自动采集系统的准确性要求。A camera with UBS interface and CMOS chip and Video for Window（VFW）technology were used to develop a computer vision system to obtain the image of beef ribeye section for acquiring grading information.The measuring and analysis methods of the grading features were introduced.The features,such as ribeye area,backfat thickness,marble distribution,muscle and fat colors were extracted by computer image analysis.Four parameters that represent the grading of beef marbling were determined.The accuracy（mean±standard deviation）of backfat thickness was（92.72±4.75）%,and that of ribeye area was（94.84±2.52）%.Assessment of grading of beef marbling was performed with back-propagation neural network,and the mean accuracy was 94.74,which meet the practical accurate requirement of computer vision system.

关 键 词：食品加工技术 检测系统 机器视觉 牛胴体眼肌切面分级 

分 类 号：TS251.7[轻工技术与工程—农产品加工及贮藏工程]