Distinction and Recognition of the 'Black Pearl' Fresh Corn Origin Based on Electronic Nose and BP Neural Network

Abstract

: 'Black pearl' fresh corns from different regions were analyzed using an electronic nose to capture the aroma profile. Principal component analysis (PCA) and discriminant function analysis (DFA) were used for multivariate statistical analysis of 200 data from two regions. Based on this, the judgment model of samples from Heilongjiang production area was built using a soft independent modeling class analysis (SIMCA) algorithm, and a back propagation neural network model was established by Pytorch software to identify and differentiate samples from different regions. The results illustrated that, although the volatile flavor of 'black pearl' fresh corns from different origins were similar, it also showed obvious origin characteristics. SIMCA model could effectively distinguish whether unknown samples come from Heilongjiang (the accuracy rate was 97%), while BP neural network model could predict and identify the origin of 'black pearl' fresh corns from unknown production areas, and the average accuracy rate was 99.44%. The combination of electronic nose technology and BP neural network model could accurately distinguish and identify the origin of 'black pearl' fresh corns.

Keywords

• fresh corn
• electronic nose
• principal component analysis (pca)
• soft independent modeling analysis (simca)
• back propagation neural network
• origin identification