Study of progress on application of hyperspectral imaging combined with deep learning approaches in detecting foods content

Predicting food component content is one of the important research directions in the field of food science and engineering, and it is also an important indicator for evaluating food quality. With the increasing attention to food quality and safety issues, the prediction of food component content is of great significance for quality control in the food industry and food safety monitoring. Traditional prediction of food component content is usually carried out through chemical analysis and physical measurement techniques, which have disadvantages such as long detection times and high costs, and may be influenced by environmental factors and sample variability. They cannot meet the requirements for non-destructive, rapid prediction of multidimensional food component content. Therefore, the development of non-destructive, rapid, and multidimensional detection methods is crucial. Hyperspectral imaging technology is a non-destructive, rapid, and high-precision technology that can obtain a large amount of spectral information from food. Deep learning, as a powerful machine learning method, has the ability to handle large-scale data and extract complex features. With the development of hyperspectral imaging technology and deep learning, an increasing number of studies are combining the two for application in non-destructive food detection. This paper reviews the application of hyperspectral imaging technology combined with deep learning in food quality analysis, including nutrient analysis, traceability identification and maturity assessment. Secondly, the research progress of hyperspectral imaging technology combined with deep learning algorithms in data preprocessing, model building and evaluation is summarized to provide reference for improving the accuracy and efficiency of food analysis.