AI-based modified atmosphere packaging for fruits and vegetables preservation: research progress and prospects

Modified atmosphere packaging (MAP) has been extensively applied in the preservation of fruits and vegetables (F&Vs). However, challenges arise from the variety of packaging materials, complex gas compositions, and diverse respiration patterns of F&Vs. Traditional mathematical tools struggle to accurately design, detect, monitor, and predict MAP for foods. Artificial intelligence (AI), as one of the versatile tools which could simulate, extend, and expand human intelligence, has demonstrated its role in many fields, as well as in MAP for F&Vs. This review first revealed the literature and research team overview of AI in the field of MAP for F&Vs through bibliometric analysis. Then, the respective classifications and joint applications of MAP and AI for F&Vs were reviewed. At present, the application of AI-based MAP in exploring mechanisms, packaging design, parameter optimization, quality monitoring, and shelf-life prediction for F&Vs has made preliminary progress. In the future, it needs to develop toward high precision, high throughput, automation, and cost-effectiveness. Meanwhile, challenges remain, including data scarcity, complex models, high learning costs, and difficulties in verifying the authenticity of AI outputs. Additionally, specific issues related to F&V MAP, such as sample variability, packaging material selectivity, and interactions between gases, samples, and films, need further attention.