一种猪肌内脂肪含量测定的方法及其改进

An Improved Algorithm of Pig's Intramuscular Fat Detection

为了解决猪肌内脂肪含量的检测,比较各类被用于检测肌内脂肪含量的方法,包括物理化学和计算机视觉等解决方法。根据比较得出的结论,提出在活体猪的肌内脂肪含量的检测中,应用超声系统将是一个很有效的办法。描述一种经过改进的基于超声图像参数分析的猪肌内脂肪含量的检测算法。算法利用超声定制探头采集猪肋骨的图像,通过对图像的频谱、直方图、梯度与纹理参数进行综合分析,建立PIMF的回归分析模型。同时也建立一个基于反馈学习的参数优化系统,因此该算法提高系统对不同猪种的适应性以及检测结果的正确性。

To solve the problem of detection of pig＇s IMF （Intramuscular Fat） which is an important topic and getting more and more concern in recent years. Many methods have been compared contains either physically or in computer vision. According to the comparison, proposes that ultrasound is a better choice to detect IMF in living swine. Describes an improved algorithm which used ultrasound system to get pig＇s image of herbs, analyzes spectrum, histogram, gradient and texture parameters of images, builds regression model of PIMF. Builds an optimized system based on feedback learning. Improves algorithm adaptability and accuracy when the method used on different breeds of pig.