Method for synchronously and rapidly detecting multiple active components in peanut kernels based on spectrum fusion

Abstract

The invention discloses a method for synchronously and rapidly detecting multiple active components in peanut kernels based on spectrum fusion, which comprises the following steps: collecting near infrared spectrum and infrared spectrum of peanut kernel samples, splicing the near infrared spectrum and the infrared spectrum to form an original fusion spectrum, preprocessing the original fusion spectrum to obtain a preprocessed fusion spectrum; determining the contents of resveratrol, folic acid and sitosterol in the peanut kernel sample by adopting a high performance liquid chromatography; based on the preprocessed fusion spectrum and the content of resveratrol, folic acid and sitosterol in the peanut kernel sample, establishing a multi-output regression model by adopting a least square support vector machine algorithm; and inputting the pretreatment fusion spectrum of a peanut kernel sample to be detected into the multi-output regression model, and synchronously outputting content predicted values of resveratrol, folic acid and sitosterol.

Images