摘要
传统猪肉等级评价及测量手段复杂低效,且易造成人为误差。本研究利用图像处理技术估算猪肉等级,实现基于计算机视觉技术的快速无损测量。本实验采用Sony717数码相机在暗箱中固定物距和焦距拍摄猪的眼肌图像,利用图像处理技术从中提取特征参数,并研究特征参数与实测眼肌面积、肌内脂肪含量及大理石花纹评分的关系。结果表明,利用图像参数与胴体实测值建立的关系模型有较好的线性关系,图像处理技术在快速无损测量猪的眼肌面积和肌内脂肪含量以及估算大理石纹评分方面表现出了巨大潜力。
The disadvantages of measure treatments before are complex and inefficient; meanwhile they have many errors made by men. The study is to predict important parameters in pork grade by Image Processing Techniques, and try to achieve quick non-vitiable measure treatment based on computer vision.We pictured loin-eye photos in a camera obscura by Sony 717 digital camera, which the focus and the distance are fixed. Then we got the eigenvalues from the pictures by Image Processing Techniques, and we built the relations among the eigenvalues and real loin-eye area, fat content in loin-eye and marble evaluation in statistical software.
出处
《肉类研究》
2004年第3期41-44,共4页
Meat Research
