基于近红外光谱的猪肉新鲜度无损检测 方法的改进

Quantitative evaluation of pork freshness based on near-infrared spectroscopy

中文关键词:  近红外光谱法  新鲜度  挥发性盐基氮  岭回归

英文关键词:near-infrared spectroscopy  freshness  total volatile basic nitrogen  ridge regression

中文摘要:

      目的 采用可见/近红外光谱技术, 结合岭回归偏最小二乘对猪肉新鲜度进行定量分析。方法 利用自行搭建的可见/近红外光谱检测系统, 采集62个猪肉样品表面380~900 nm范围内的反射光谱数据, 进行标准正态变量变换(standard normal variable transform, SNVT)预处理后, 建立偏最小二乘(partial least square regression, PLSR)模型。利用模拟退火算法(simulated annealing, SA)和粒子群算法(particle swarm optimization, PSO)进行岭参数寻优, 建立猪肉挥发性盐基氮(total volatile basic nitrogen, TVB-N)的岭回归模型。结果 所建模型的相关系数和误差分别为0.9819、1.2785 mg/100 g和0.9781、1.4628 mg/100 g。结论 所建立的模型取得了较好的结果, 利用岭回归偏最小二乘实现了对最小二乘估计的改良, 更加验证了可见近红外光谱技术对猪肉新鲜度进行定量分析的巨大应用潜力。

英文摘要:

      Objective To quantitatively evaluate pork freshness based on visible-near infrared spectroscopy combined with ridge regression. Methods A laboratory visible/near-infrared spectroscopy system was built to collect reflectance spectra in the 380～900 nm in 62 pork samples and the ridge regression was employed to establish partial least squares regression (PLSR) models. The reflectance spectra of samples were performed with pretreatment method of standard normal variable transform (SNVT), and then PLSR model was built. Simulated annealing (SA) algorithm and particle swarm optimization (PSO) algorithm were employed to optimize ridge parameter k. Then the ridge regression model for total volatile basic nitrogen (TVB-N) value was established. Results The correlation coefficients and errors were 0.9819 and 1.2785 mg/100 g, and 0.9781 and 1.4628 mg/100 g, respectively. Conclusion The models built had achieved good results, and the ridge regression using partial least squares could improve the model performance of traditional partial least squares models. It verified the huge potential application of visible near infrared spectroscopy to quantitatively analyze pork freshness