近红外技术快速检测冷鲜羊肉品质的研究

Rapid evaluation of mutton quality by near-infrared spectroscopy

中文关键词:  近红外光谱法  偏最小二乘法  羊肉品质  评定

英文关键词:near-infrared spectroscopy  partial least squares  mutton quality  evaluation

中文摘要:

      目的 建立近红外反射光谱法, 结合偏最小二乘回归法对冷鲜羊肉进行快速无损检测。方法 从农贸市场购买新冷鲜羊肉100批, 通过分析一阶导和二阶导, S-G平滑(savitzky-golay, S-G)和Norris平滑方式建立羊肉的近红外光谱参数与样品的蛋白质、脂肪、水分含量之间的对应关系模型。结果 蛋白质含量最优模型选择5981.72~ 5961.74 cm?1波段, 二阶导, S-G平滑方式, 相关系数为0.9312, 建模方差为0.803; 脂肪含量最优模型选择5207~7362 cm?1波段, 一阶导, Norris平滑方式, 其相关系数0.9157, 建模方差为4.703; 水分含量最优模型选择6800~9000 cm?1波段, 一阶导, Norris平滑方式, 相关系数为0.9200, 建模方差为0.039。结论 所建立模型的精确度符合检测要求, 为羊肉品质的分析测定提供了一种快速方便的检测方法。

英文摘要:

      Objective To establish a method for quickly nondestructive testing of cold fresh mutton by visible-near infrared (NIR) spectroscopy with partial least square regression. Methods One hundred batches of cold fresh mutton were purchased from farmers' markets, the corresponding relation models between the near-infrared spectral parameters of mutton and the protein, fat and water content of samples were established by analyzing the first derivative and second derivative and using S-G smoothing and Norris smoothing methods. Results The model of protein content was optimized by selecting 5981.72-5961.74 cm?1 band, second derivative, S-G smoothing method, and the calibration coefficient of protein content was 0.9312, with the variance of 0.803. The model of fat content was optimized by selecting 5207-7362 cm?1 band, first derivative, Norris smoothing method, and the calibration coefficient of protein content was 0.9157, with the variance of 4.703. The model of water content was optimized by selecting 6800-9000 cm?1 band, irst derivative, Norris smoothing method, and the calibration coefficient of protein content was 0.9200, with the variances of 0.039. Conclusion The accuracy of the model is conformed to the testing requirement, and it provides a quick and convenient method for the analysis and determination of mutton quality.