A Machine Learning Enhanced Portable Electronic Nose for Seafood Freshness Evaluation

In this paper, we have proposed a small model to determine the freshness of seafood. The seafood, when spoilt, releases certain gases. So, we have used MQ series gas sensors like MQ2, MQ4, MQ8 and MQ135 to detect the gases. We also used a DHT11 sensor because the temperature and humidity change the spoiling rate. All the readings are sent to an Arduino Uno, and whenever the gas values cross the predefined limit, the system sends a message to the user through the GSM module and also alerts the user through a buzzer. A random forest machine learning algorithm is used to classify the seafood fresh and spoiled conditions. We trained the algorithm using the data collected from sensors while testing. From the tests conducted, the system response was found to be quick and steady. Since the components are simple and low cost, this model can be used in homes and small food storage places.