| Popis: |
Food wastage dew to spoilage is one of the key global problems, the quantity of food loses and spoilage per year is 40-50% for root crops, fruits, and vegetables. This paper proposes an edge IoT & machine learning based approach for food quality monitoring. The need for this type of system is to correctly classify the vegetables and fruits in three categories fresh, semi-fresh & spoiled, further to alert about environmental conditions, either it's favorable or not using the expert algorithm developed under particular standard environmental data. The proposed system employed with Raspberry Pi as the main processing unit, temperature & humidity sensors along with three VoC gas sensors. To classify the vegetable and fruit, the sensed data (Gas Sensors MQ2, MQ3 & MQ9 along with humidity & temperature sensor) were collected over a week & machine learning models Logistic Regression, Linear SVC, rbf SVM Classifier, Random Forest with the accuracy of 99.1%, 99.9%, 99.96% & 99.96% respectively have been used to classify the data. Further, the classified and processed result was uploaded on the firebase cloud to make it possible to be monitored from anywhere in the world using a particular developed android application. The uniqueness of the system is, it uses edge IoT and its software is programmable so that it can be used to preserve any type of food items by just filling up some credentials in the settings section of GUI. |
