Non-destructive Detection of Codling Moth (.Cydia pomonella L) Damage in Apple Fruit Using Hyperspectral Imaging Method

In this study, Hyperspectral Imaging method in the range of 400-1000nm has been applied to detect infested apples from normal ones. After preparing the infested samples acording to standard and transfering the samples to the lab, the images were taken under controled situation. Afterwards, average relative reflectance was extracted from the region of interest and then was pre-processed. Finaly the average relative reflectance data was classifeid using different machine learning methods including Discriminant Analysis (DA), K-nearest neighbor (KNN) and Decision Tree (DT) techniques. Results showed that classification of infested samples from normal ones was possible with the classification rates of 96% and 94% for normal and infested apples, respectively. The highest classification rate achieved for DA method. Also, the optimum wavelengths were extracted from the spectrum in order to develop Multispectral Imaging system. The results of this research indicate the high performabce of Hyperspectral Imaging Method for non-destructive detection of infested samples for application in apple grading machines.