Detection of Contaminated Hazelnuts under UV Illumination

Hazelnut (Corylus avellana L.) is one of the most important crops in the world. Due to its high nutritional composition and enhanced flavors, it is used in numerous confectioneries, baking, and chocolate industries. During various cultivation, harvesting, and storage phases, hazelnuts can be prone to bacterial and fungal infections which damage their nutritional values as well as result in bad odor and unpleasant taste. This work proposes a solution to detect rotten hazelnuts by means of Ultraviolet (UV) light. In order to uniformly illuminate hazelnut samples and to capture high-quality images, an axial illuminator is specially modeled and printed using a 3D-printer. The sample images are captured at 450 nm and 750 nm. The healthy and rotten hazelnuts behave differently in the presence of UV light, such that the healthy parts produce fluorescence after absorbing the UV light and appear brighter, whereas the rotten hazelnut samples do not produce any fluorescence and are therefore darker in the captured images. The two images are then combined by using image division operation with suitable scaling factors. The resulting gray-scale images are converted into binary images using image thresholding. These binary images represent the damaged or rotten hazelnuts as black and the healthy hazelnuts are white in color. The percentages of black and white pixels determine whether a hazelnut sample belongs to the healthy or the rotten class