Detection and Classification of Bruises on ‘Red Delicious’ Apples Using Active Thermography

Identifying and sorting healthy agricultural products from damaged products reduces casualties and losses caused by the spread of disease to unhealthy samples. Due to Iran's insignificant contribution to agricultural exports, the use of non-destructive methods, such as thermography in the sorting and grading of the fruit, is necessary. In this research, active thermography method was used to identify the Red delicious cultivars of apple. In order to study the temperature variations of apples, used a factorial experiment was conducted in a randomized complete design with three independent variables including bruises, heating temperature and cooling time. The results of variance analysis of surface temperature of healthy and bruise apples showed that the levels of bruises, heating temperature and cooling time on the surface temperature of bruises apples had a significant effect at 1% probability. Statistical and texture features were extracted from thermal images and the artificial neural network method was used to classify two healthy classes and bruised. Neural network accuracy with a hidden layer and 15 neurons was obtained 100%. The results indicate that the difference in temperature of the bruised and sound tissue can be considered as a criterion for the classification of apples. Also, Active thermography is an efficient and high-tech method for detecting apple bruises.