基于近红外光谱的鸡蛋产地溯源

为了快速、准确、无损地追溯鸡蛋的不同产地,借助于近红外光谱技术,采用主成分析结合PLS-DA判别模型和簇类独立软模式法(SIMCA)建立了鸡蛋的溯源模型。利用标准正态变量(standard normal variate,SNV)、Savitzky-Golay平滑滤波(SG)和多元散射校正(multiplicative scatter correction,MSC)等方法对原始光谱数据进行了预处理,结果表明SG(3点)平滑处理结果最好;利用主成分分析方法对不同地区的鸡蛋进行聚类分析,发现当主成分数为3时,建立的SIMCA溯源效果最好。结果表明,在显著水平0.05时,4个地区(朔州、吕梁、太谷、运城)验证集的识别率均为100%,其中吕梁和运城地区的拒绝率为100%,朔州和太谷地区的拒绝率为98.6%。说明SIMCA模式建立的模型基本能够判别鸡蛋产地。

The Origin of Eggs Based on Near Infrared Spectroscopy

In order to analyze the different production sites of eggs efficiently, accurately and nondestructively, the near-infrared spectroscopy technology was used. The PLS-DA judgment model and SIMCA were combined in the analysis process, the traceability model of eggs could be effectively established. More importantly, the original spectral data was processed in advance by means of standard normal variate, SNV, savitzky-golay, multiplicative scatter correction, MSC, etc., and the smoothing treatment in SG (3 points) was the optimal experimental result. In addition, the principal component analysis method was used in the experiment, which could be used for clustering analysis of eggs in different areas. It was found that when the principal component fraction was 3, the established SIMCA had the best traceability effect. In addition, the results of analysis data showed that when the significance level was 0.05, the recognition rate of the four regions (shuozhou, luliang, taigu and yuncheng) was 100%, of which the rejection rate of luliang and yuncheng was 100%, and that of shuozhou and taigu was 98.6%. These experimental data show that the model established by SIMCA model can effectively identify the origin of eggs under certain conditions.