Investigating the Effect of Oxygen, Carbon Dioxide, and Ethylene Gases on Khasi Mandarin’ Orange Fruit during Storage

Abstract

This study presents on predicting the shelf life of’Khasi mandarin’ oranges stored under specific conditions through the analysis of their respiration rate and ripeness levels. By employing a finely tuned deep convolutional neural network (CNN) trained on 1284 images of’Khasi mandarin’ oranges, the research classifies the fruit into four ripeness categories: unripe, partially ripe, ripe, and over-ripe. Stored at temperature (26.39 ± 3.07 °C) and humidity level between 60 and 80%, the CO₂ respiration rate (RR_CO2) was calculated based on enzyme kinetics principles to correlate with these ripeness levels, indicating a shift toward anaerobic respiration as the fruit undergoes ripening and metabolic changes. Moreover, ethylene release, initially at 0.43 ± 0.017 mL/kg/h on day 0, precipitously increased to 6.943 ± 0.0296 mL/kg/h by day 17, reflecting the ripening process. A support vector regression model predicts shelf life and ripeness levels, creating an AI-based soft sensor applicable to various fruits. This approach enables dynamic decision-making in pricing, logistics, and storage conditions, reducing fruit waste and economic losses. Integrating AI-driven solutions into postharvest handling enhances efficiency and sustainability in fruit distribution and storage, benefiting agricultural and retail industries.