【摘要】 [目的]研究在无损情况下猪肌内脂肪含量的估测方法，以提高预测模型的准确性。[方法]以229份硒都黑猪性能测定数据及超声波图像数据为研究对象，利用BioSoft Toolbox For Swine软件中感兴趣区域（region of interest,ROI）模块分割超声波图像，确定最佳感兴趣区域像素框大小（ROI Size）和个数，并通过逐步回归分析，以猪肌内脂肪含量为因变量，图像读取参数、体尺指标为自变量构建猪活体肌内脂肪含量估测模型。[结果]肌内脂肪含量与图像读取参数、宰前活重、胸围、腹围具有显著相关性，运用逐步回归分析得到脂肪含量估测模型，IMF=0.2495+0.1752×P+0.0070×W-0.0055×B+0.0159×A，模型拟合度R~2为0.725 2。[结论]运用超声波技术可以较准确地测定猪活体肌内脂肪含量，为实现种猪性能改良提供了技术支撑。更多还原

【Abstract】 [Objectives] Study the method of evaluating the content of intramuscular fat in pigs under non-destructive conditions to improve the accuracy of predictive models. [Methods] The performance measurement data and ultrasonic image data of 229 Xidu black pigs were used. The Region of interest(ROI) module in BioSoft ToolboxFor Swine software was used to segment ultrasonic images to determine the optimal ROI size and number. The model for in vivo evaluating the content of intramuscular fat in pigs was constructed with a stepwise regression analysis using the content of intramuscular fat in pigs as the dependent variable, parameters of image reading and indicators of body size as independent variables.[Results] There was a significant correlation between the content of intramuscular fat and parameters of image reading, live weight before slaughter, chest circumference, and abdominal circumference. A model for evaluating the content of intramuscular fat in pigs obtained was expressed as IMF=0.2495+0.1752×P+0.0070×W-0.0055×B+0.0159×A, with R~2 of 0.725 2. [Conclusions] The use of ultrasound technology can accurately determine the content of intramuscular fat in live pigs. It will provide technical support for improving the performance of breeding pigs.更多还原