Non‐destructive prediction of TVB‐N using color‐texture features of UV‐induced fluorescence image for freeze–thaw treated frozen‐whole‐round tilapia

Abstract

BACKGROUNDThe investigation of ultraviolet (UV)‐induced fluorescence imaging coupled with machine learning was conducted to non‐destructively detect the total volatile basic nitrogen (TVB‐N) of frozen‐whole‐round tilapia (FWRT) during freezing and thawing (F‐T). The UV‐induced fluorescence images of FWRT at the wavelength of 365nm were acquired by self‐developed fluorescence image acquisition system. A total of 169 color and texture features based on RGB, HSI, and L*a*b* color spaces and gray level co‐occurrence matrix (GLCM) were extracted, respectively. Successive projections algorithm (SPA) was employed to select the optimal 16 features to achieve feature dimension reduction modeling. With full and extracted features as input, the models of partial least squares regression (PLSR), least‐squares support vector machine (LS‐SVM), and convolutional neural network (CNN) were established for TVB‐N prediction.RESULTResults indicated that the full features‐based CNN performed better than SPA based prediction models (SPA‐PLSR and SPA‐LSSVM). CNN model was determined to be the optimal with RP2 value of 0.9779, RMSEP value of 1.1502×10‐2 g N/kg, RPD value of 6.721 for TVB‐N content predicting.CONCLUSIONIt indicated that the CNN method based on UV fluorescence imaging technology has potential for quality and safety detection of FWRT.This article is protected by copyright. All rights reserved.