Banana maturation is typically categorized into seven stages, labeled from class 1 (unripe green fruit) to class 7 (ripe banana, completely yellow with brown spots, technically named “tiger”) . This classification is particularly important for logistics: bananas are harvested unripe (class 1), shipped to a destination country, ripened using ethylene, and finally distributed at maturation stages of 3 or 4. Specifically, bananas at stage 3 are destined for distant markets, while stage 4 are for local markets. However, this standard approach is flawed as it overlooks the intermediate maturation stages between 3 and 4, potentially leading to significant wastage at both retail and consumer levels. The present study aims to accurately identify the maturation stage of Dwarf Cavendish bananas (Musa acuminata L.) using computer vision technologies. Specifically, chemical and physical analyses were performed for identifying intermediate stages between 3 and 4 by using k-means clustering. Two intermediate (INT) classes were identified, leading to a total of four classes studied: 3, INT1, INT2 and 4. Subsequently, two computer vision techniques were applied to recognize the four maturity stages: i) longitudinal profile colorimetric analysis; and ii) tonal distribution analysis. The experimentation was conducted on a total of 300 fruits. Classification models were developed using Partial Least Squares Discriminant Analysis. The best predictive model was achieved using the longitudinal profile colorimetric analysis dataset. Cross-validation classification errors were as follows: class 3 1.46%; class INT1 2.43%; class INT2 0.49%; and class 4 0.49%.