Non-destructive prediction of total volatile basic nitrogen (TVB-N) content of Litopenaeus vannamei using A bi-channel data acquisition of Colorimetric sensing array

The change of volatile organic compounds (VOCs) caused by spoilage bacteria are typical characteristics of seafood decay. This study investigated the freshness of shrimp (Litopenaeus vannamei) using a bi-channel data acquisition based on the colorimetric sensing array (CSA) technique. First, nine color-sensitive dyes were selected to capture VOCs changes during shrimp spoilage. Then, both image and spectral channel data from the CSA were used to predict the total volatile basic nitrogen (TVB-N) content and evaluate its prediction ability, respectively. The original full data from the spectral channel performs better than the image channel. Next, four optimization algorithms are tried to filter spectral feature variables. Finally, the variable combination global analysis-iterative retained information variables (VCPA-IRIV) algorithm combined with a partial least squares (PLS) model was determined for predicting the TVB-N of shrimp, achieving the best precision and robustness performance with the Rp2, RMSEP, and RPD were 0.9734, 1.54, and 6.14, respectively. Therefore, this study provides a new approach to evaluate shrimp freshness rapidly.