Meat spoilage characterization via electronic nose

Abstract

Meat is one of the food varieties that spoils rapidly when not chilled. Meat spoilage in public markets has been one of the main contributors to food wastage and imposes health hazards when consumed. Although the meat vendors intend to sell fresh and good products, they cannot identify how much time meat will reach spoilage, much more the buyers. A system that can characterize meat samples to predict spoilage time will benefit not only the vendors and regulation authorities but moreover the consumers. This study aims to develop an electronic nose system for characterizing the gases emitted when meat is approaching spoilage. The system will look for patterns in the levels of methane (CH₄), ammonia (NH₃), and hydrogen sulfide (H₂S). An array of metal oxide semiconductor (MOS) sensors placed in an air -tight chamber measures the emitted gas particles by the meat samples. In each experiment, a meat sample is placed in the enclosure while monitoring the concentration of CH₄, NH₃, and H₂S at different time intervals. Several variables affect the rate of spoilage of meat. The meat samples’ weight (or volume) and the time elapsed at room temperature are used as the independent variables. The level of the concentration of CH₄, NH₃, and H₂S at different time intervals becomes the feature vectors for training different classifiers. Four classifiers were considered for meat spoilage characterization but the weighted KNN performed best with an accuracy of 98.33%. The system has successfully detected spoilage in meat as well as the experts in the field do. Spoilage can be detected by taking into account the gas concentration levels of CH₄, NH₃, and H₂S. In future works, temperature effects can also be integrated with the system.