Non-Destructive Sensing and Intelligent Quality Prediction During Fruit Drying: From Quality Formation to Decision Support.

Fruit drying transforms a living, water-rich tissue into a stable food through coupled changes in moisture distribution, structure, color, nutrients, and aroma. Although drying technologies and non-destructive sensing have advanced rapidly, these fields have largely developed in parallel, leaving the relationship between quality formation and measurable process signals insufficiently resolved. Here, physical and chemical changes during drying are connected to the signals that can support quality prediction. Current evidence shows that moisture loss and surface appearance are the most tractable real-time targets. Texture, bioactive retention, and flavor remain less accessible because their signals depend more strongly on internal structure, reference chemistry, or sensory response. Optical, magnetic-resonance, thermal, volatile-sensing, and electrical approaches consequently provide complementary rather than interchangeable views of the product. Multimodal models improve prediction when the added signals resolve different aspects of drying, but redundant inputs can increase complexity without improving transferability. Progress toward intelligent fruit drying therefore depends on matching sensors to the evolving product state, validating models beyond individual batches and instruments, and linking predictions to practical process decisions. This process-quality perspective provides a basis for moving from retrospective quality assessment toward reliable monitoring and controlled drying.