摘要
利用近红外光谱技术建立快速测定鱼油中EPA和DHA含量的方法。分别测定EPA和DHA的含量以及收集鱼油光谱信息,比对了多种单一和复合预处理方法以及数学建模方法。结果表明:Normalize-SNV和Detrend-MSC分别为EPA和DHA预测模型最佳的预处理方法,PLS为最佳的建模方法,其中EPA预测模型R_C为0.966,SEC为1.960,DHA预测模型RC为0.945,SEC为1.281。利用近红外光谱技术可以快速测定鱼油中EPA和DHA含量。
The method of rapid determination of EPA and DHA contents in fish oil was established by NIR technology. The EPA and DHA contents were determined and the spectral information of fish oil was collected. The different single and composite pretreatment methods and mathematical modeling methods were compared. The results showed that the Normalize-SNV and Detrend-MSC were the best pretreatment methods for EPA and DHA,respectively. PLS was the best modeling method. The R_C and SEC of the EPA prediction model were 0. 966 and 1. 960 respectively,and the R_C and SEC of the DHA prediction model were 0. 945 and 1. 281 respectively. The EPA and DHA contents in fish oil could be determined rapidly by NIR technology.
出处
《中国油脂》
CAS
CSCD
北大核心
2017年第10期138-142,共5页
China Oils and Fats
基金
中国农业科学院科技创新工程(CAAS-ASTIP-201X-IAPPST)
