DETECTION OF BRUISES IN MAGNETIC-RESONANCE IMAGES OF APPLES

A method for fast, computerized detection of bruises in magnetic reson ance images (MRI) of apples was developed. It uses a simple thresholdi ng technique combined with considerations of the apple geometry to dis tinguish between bright pixels representing the vascular system, and t hose representing bruises. The algorithm was tested on 20 images of te n Jonathan apples (before and after bruising), ten images of Golden De licious apples and ten images of Hermon apples, and it demonstrated it s ability to distinguish between bruised and non-bruised apples. This is the first step in the development of a sensing method using three-d imensional imaging and image analysis for detection of bruises.