生鲜猪肉细菌总数的高光谱特征参数研究

Assessment of characteristic parameters of total viable count of fresh pork based on hyperspectral images

中文关键词:  生鲜猪肉  高光谱  细菌总数  洛伦兹拟合  多元线性回归

英文关键词:pork  hyperspectral image  bacteria total viable counts  Lorentz fitting  multi-linear regression

中文摘要:

      目的 通过对高光谱数据进行洛伦兹拟合参数的分析, 讨论高光谱技术对生鲜猪肉细菌总数预测的可行性。方法 63个猪肉样品贮存于4 ℃冰箱中, 每天随机取出4块样品, 在400~1100 nm波长范围内获取猪肉表面的高光谱散射图像, 从高光谱图像中提取猪肉的反射光谱曲线, 利用洛伦兹函数进行拟合, 然后用单参数和不同参数结合的方法建立多元线性回归模型。结果 多参数结合的方法比单个参数建立的模型更好, 最好的模型结果是三个参数结合建立模型, 校正集相关系数为0.96, 标准差为0.42; 预测集相关系数为0.89, 标准差为0.46。结论 利用高光谱成像技术结合洛伦兹函数对快速检测猪肉细菌总数具有一定的可行性。

英文摘要:

      Objective To develop a new technique by using Lorentz fitting parameter analysis so as to discuss the feasibility of predicting the bacteria total viable count (TVC) of fresh pork by using hyperspectral imaging technology. Methods Totally 63 fresh pork samples stored at 4°C were used in the experiment and 4 samples were taken out randomly each day. Hyperspectral scattering images and spectral reflectance optical data were acquired from surface of pork samples in the region of 400~1100 nm. Lorentz function was applied to fit the scattering profiles of pork. Both individual parameters and integrated parameters were explored to develop the multi-linear regression models for predicting pork TVC. Results The results indicated that the integrated parameters could perform better than individual Lorentz parameter. The best result for predicting pork TVC was achieved by the form of (a, b, c), with the correlation coefficient and standard error of 0.96 and 0.42 for calibration set, and 0.89 and 0.46 for prediction set, respectively. Conclusion Hyperspectral scattering technique combined with Lorentz function is potential for rapid determination of pork TVC.