Abstract:
The quality of edible fruits often depends on their freshness; there is no acceptable mechanism to check the freshness of fruits. A combination of technologies such as Arduino and sensors can detect the freshness of the fruits. Oxygen (O2) and carbon dioxide (CO2) gases, and humidity are related to the freshness of the fruits. Thus, this research aims to predict the freshness of fruits by observing CO2 release, O2 absorption and water vapour release after harvesting. Papaya and watermelon were selected for this study, and these fruits were categorized into three groups (0.5-1 kg, 1-1.5 kg, and 1.5-2 kg). After the harvest, three freshness factors (CO2, O2, and humidity) were measured at the intervals of one and three days and after the first and second weeks. A closed system consisting of CO2 and O2 sensors, and a humidity sensor was set up to detect the changes of the above factors of the fruits. Then, a supervised machine learning model was developed using a logistic regression algorithm to predict the freshness of fruits. The collected sensor data was used to train the machine learning model. After entering fruit type, weight, a difference of oxygen and water-vapour constatation as inputs for the model, the model will predict the freshness of the fruit as a percentage. Analyzed results showed, the rate of O2 absorption gradually increases after harvesting, and water-vapour release gradually decreases. However, it is impossible to get an accurate CO2 value due to the low sensitivity of the sensor used. Due to the low sensitivity of the sensor used in this research, it took a longer duration (>45 minutes) to obtain significant changes in the factors. It is recommended to use sensors with higher sensitivity for better detection abilities fruit freshness.
