Banana (Musa spp.) is a primary climacteric fruit characterized by a rapid surge in ethylene production and respiration post-harvest. Accurate ripeness detection is critical for supply chain management, export logistics, and reducing global food waste, which accounts for nearly 30% of harvested bananas. Traditional methods-visual inspection and destructive chemical testing e.g., Total Soluble Solids (TSS) and Titratable Acidity (TA)-are subjective and labor-intensive. Recent years have seen a paradigm shift toward non-destructive testing (NDT) powered by Deep Learning (DL), Computer Vision (CV), and multi-modal sensor fusion. This comprehensive review critically examines diverse methodologies employed for identifying banana ripening stages, including dielectric properties, deep learning, artificial intelligence and neural networks, image processing, laser-induced backscattering imaging, and spectroscopy. We evaluate the underlying principles, effectiveness, current limitations, and practical applicability of each approach. The review highlights the challenges associated with standardizing ripeness identification across the popular banana cultivars and environmental conditions, as well as the computational and practical hurdles of advanced technologies. Finally, we discuss emerging trends and propose future research directions, emphasizing the integration of multi-sensory data and advanced computational models for developing robust, cost-effective, and scalable solutions that enhance sustainable post-harvest management and reduce food waste.
Renugadevi et al. (Mon,) studied this question.