Fruit Defect Inspection System Using Image Processing and IoT Framework

The consumption of fruits is in high demand due to their nutritional value. Most of the fruits available in the market are processed through chemical practices hampering their quality. Exposure to fruit preservatives and carbides enhances its life and ripens it faster. However, eating such fruit results in poor health and increases the probability of getting infected by various threatening diseases such as cancer, tuberculosis, etc. Organic farming is practiced in some areas of India to achieve fruit quality, but its inadequate to fulfill the demand. To overcome the issues mentioned, a model based on IoT is proposed in this research. A system to separate quality fruits from a basket is presented in this article. The classification will be done using a deep learning technology, Convolutional Neural Network (CNN) which uses a database consisting of pictures of three fruits particularly, apples, oranges, and bananas are used in the experiment. The input from the alcohol (MQ3) sensor and methane (MQ4) sensor are fed forward to node MCU. The input in turn is provided to Arduino UNO for comparison with preprocessed audit set. The inception V3 algorithm is used for classification purposes. This research proposes a cost-effective and near-to-accurate solution to issues in automated fruit quality identification.