A computer vision based automatic sorting system for apple fruits is introduced. The fruit area is extracted from the
background and it is eroded to reduce undesired effects of illumination. After this preprocessing, defected areas of fruit
are segmented by an artificial neural network classifier