基于VIS/NIR高光谱散射特征预测牛肉的嫩度  

Prediction of beef tenderness using VIS/NIR hyperspectral scattering characteristics

作　　者：吴建虎[1] 彭彦昆[1] 高晓东[1] 陈菁菁[1] 黄慧[1] 

机构地区：[1]中国农业大学工学院,北京100083

出　　处：《食品安全质量检测学报》2009年第1期20-26,共7页Journal of Food Safety and Quality

基　　金：国家自然科学基金(30771244);北京市自然科学基金(6082016);国家863高技术研究发展计划(2008AA10Z210)资助项目

摘　　要：本文研究利用VIR/NIR光谱散射特征预测成熟 7 天牛肉的嫩度.开发高光谱散射成像系统,获取新鲜牛肉 400～1100 nm 波长范围高光谱散射图像,对牛肉嫩度进行预测和分级.利用洛伦兹函数,拟合各个波长处的散射曲线,获取不同波长散射曲线的洛伦兹分布函数参数.使用逐步回归方法,选择最佳波长及相应的拟合参数,建立线性回归模型预测牛肉的嫩度,使用全交叉验证方法评价模型的性能.使用散射曲线的峰值建立的模型对嫩度的预测结果最好,预测相关系数为0.86,预测残差为11.7 N.以嫩度剪切力值 58.8 N 为界将牛肉分为粗糙牛肉组和嫩牛肉组,对嫩度的分级准确率是 91%.该研究表明,利用牛肉的散射特征可以对牛肉嫩度预测和分级.Hyperspectral scattering images were captured from 2-day raw beef to predict 7-days aged tenderness.Fresh strip loin cuts were collected from 2-day postmortem carcass.After imaging,the samples were vacuum-packed and aged to seventh day,and then their tenderness values were measured as references.The optical scattering profiles were extracted from the hyperspectral images and fitted to the Lorentzian distribution(LD) function with their parameters.LD parameters were calculated at individual wavelength.Stepwise regression was used to determined optimal combinations of wavelengths for each of parameters.The LD parameters were then used to establish multi-linear regression(MLR) models to predict the beef tenderness respectively.The full cross validation method was used to examine the performance of models.The models were able to predict beef tenderness with the correlation coefficient of cross validation Rcv = 0.86 and SEcv=11.7 N.The results demonstrated that the scattering characteristics were useful for determination of beef tenderness.

关 键 词：牛肉嫩度 高光谱散射特征 洛伦兹分布函数 多元线性回归 

分 类 号：TS251.7[轻工技术与工程—农产品加工及贮藏工程]