Artificial Intelligence Empowered Multispectral Vision Based System for Non-Contact Monitoring of Large Yellow Croaker (Larimichthys crocea) Fillets.

Shengnan Wang,Avik Kumar Das,Peng Liang,Jie Pang

doi:10.3390/foods10061161

Abstract

A non-contact method was proposed to monitor the freshness (based on TVB-N and TBA values) of large yellow croaker fillets (Larimichthys crocea) by using a visible and near-infrared hyperspectral imaging system (400–1000 nm). In this work, the quantitative calibration models were built by using feed-forward neural networks (FNN) and partial least squares regression (PLSR). In addition, it was established that using a regression coefficient on the data can be further compressed by selecting optimal wavelengths (35 for TVB-N and 18 for TBA). The results validated that FNN has higher prediction accuracies than PLSR for both cases using full and selected reflectance spectra. Moreover, our FNN based model has showcased excellent performance even with selected reflectance spectra with rp = 0.978, R2p = 0.981, and RMSEP = 2.292 for TVB-N, and rp = 0.957, R2p = 0.916, and RMSEP = 0.341 for TBA, respectively. This optimal FNN model was then utilized for pixel-wise visualization maps of TVB-N and TBA contents in fillets.

Highlights

Large yellow croaker (Larimichthys crocea), as one of the most economically valuable marine fish in China, has a unique flavor and positive effects on health arising from its constituent proteins, polyunsaturated fatty acids, and carbohydrates, with a predicted substantial market size in south east Asia [1,2,3]
The presence of a wide range of total volatile basic nitrogen (TVB-N) and TBA indicators indicated that the calibration dataset had an excellent performance in establishing a reliable calibration model
A visible and near-infrared (VIS-NIR) hyperspectral imaging system empowered with partial least squares regression (PLSR) and forward neural networks (FNN) was conducted to rapidly and non-invasively monitor the freshness of large yellow croaker fillets

Summary

Introduction

Large yellow croaker (Larimichthys crocea), as one of the most economically valuable marine fish in China, has a unique flavor and positive effects on health arising from its constituent proteins, polyunsaturated fatty acids, and carbohydrates, with a predicted substantial market size in south east Asia [1,2,3]. Low-temperature storage technology is a common technique for preservation and extends the shelf-life of large yellow croaker products due to its convenience and low cost [4]. Lipid oxidation, and microorganism growth are inevitable activities in fish muscle post-mortem due to the high levels of nutrient and moisture content in large yellow croaker, which significantly impact freshness and consumer acceptance and reduce economic efficiency. A scalable evaluation system of accurate traits for large yellow croaker freshness is an essential process

Objectives

Results

Conclusion