Fish Quality Recognition using Electrochemical Gas Sensor Array and Neural Network

Abstract

Identification of the fish quality is needed to determine the level of freshness so that it can be consumed safely. Usually, the recognition of the fish quality through physical and odor examination by humans. This can be dangerous because spoiled fish produces poisonous gas and a pungent odor from the metabolic processes of microorganisms. This study has developed a tool for recognition of the fish quality using an electrochemical gas sensor array and a Neural Network algorithm. The electrochemical gas sensor consists of amperometric and conductometric types. This sensor data is then fed to the Neural Network algorithm which is implemented in the Arduino Due microcontroller. The experimental results show that the fish quality produces a different sensor response. The more fish decay, the greater the sensor response. This system can recognize the fish quality including fresh, half-fresh, and rotten with a success rate of 80%.