Research Information System
4 thINTERNATIONAL Conference on Advanced Engineering, Bayburt, Turkey, 28 September 2022 - 30 September 2023, no.34, pp.529-537, (Full Text)
Electronic noses that imitate the human olfactory system can make precise measurements at levels that the human nose cannot detect, thanks to the support of artificial intelligence. They are also mechanisms that enable the perception, identification and classification of odors. The perceving sensitivity of the electronic nose is higher than the human nose. This system can do the same job for years. In addition, the sensors used in the system allow to detect gases which the human nose can not perceive. For instance, it allows early detection of spoiled foods. Although the apple is an important fruit due to its high vitamin and mineral nutrient content and long storage period in the worldwide, it is susceptible to spoilage caused by fungi. Electronic nose has a significant advantage in terms of early detection of fungal spoilage and prevention of post-harvest losses. The data set used in this study, fresh apples and apples infected with Aspergillus niger, Penicillium expansum and Penicillium crustosum fungi, totally 160 ripe apples were collected in 4 different groups. The data set was divided into test and training groups. In order to determine the characteristics of apples, various feature extraction methods and wavelet transform methods were applied to apple samples in the data set. Dual classification of fungus-infected apples and fresh apples achieved 91% classification accuracy thanks to the K-nearest neighbor classification method. The results of the study showed that the deterioration of apples can be practically predicted thanks to the electronic nose.