Evaluation of physio-chemical characteristics of ‘Fard’ banana using computer vision system

After harvest, the banana is subjected to different physiological changes at ambient storage. Using image processing to analyze the quality of bananas gained a huge interest. This study was conducted to evaluate the most common physicochemical quality attributes of ‘Fard’ banana fruit during 20 d of storage at 22 °C. A mature banana fruit at the dark green stage was obtained from a farm and then stored in the laboratory at ambient temperature. Bananas' size, peel/pulp firmness, total soluble solids, acidity, pH, and sugar:acid ratio were evaluated on a daily basis. The image processing was employed to determine the R (Red), G (Green), and B (Blue), that have been converted to X (component with color information), Y (lightness), and Z (component with color information), and then to L* (lightness), a* (redness and greenness), and b* (yellowness) color space which highly used to describe the skin color of a banana.The relation between quality attributes was performed using a heatmap matrix. The kinetic models were also determined to describe the quality changes as a function of time during storage. The fruit weight loss increased by 18.20% and the transpiration rate became 0.70 mg kg⁻¹s⁻¹ on day 15 of storage. A continuous loss in the banana peel/pulp firmness was revealed. The trend of changes in the lightness and yellowness values were the same and were significantly different during the 20 d of storage. The total soluble solids correlated positively with titratable acidity (r ≥ 0.808) and sugar: acid ratio (r = 0.932) and negatively with pH (r = −0.881). The study emphasized the importance of fruit skin color using color spaces to explain other quality attributes during storage, where CIELab color space showed the best relationship with other banana quality characteristics.