DETECTION OF DEFECTS IN FRUITS USING IMAGE PROCESSING IN MATLAB

Abstract

Human Operators have been replaced by Machines and various new technologies in the field of Fruit processing. The automation of tasks done by the human operators is highly encouraged in many industrial applications. One such important task is to detect the defects in fruits based on the damage caused to it on its peel. The defect in fruits is identified mainly based upon the damage caused to its peel, when the detection is done manually. To automate this process, Image Segmentation is one of the main techniques the system’s accuracy heavily depends on as it is the first step that identifies the flaws in the fruits. The image fed to the system is converted into a digital image to either enhance it or collect some necessary information from it using Image Processing. This segmentation of fruit images can be done using various methods that have been developed over time. A hybrid algorithm is proposed in this paper which performs the segmentation using split and merge method. The K-means algorithm splits the original image based on the Euclidean colour distance in L∗a∗b∗ space to return a segmented result. Then the image is merged after a graph representation of the picture. This is an efficient procedure that can replace the manual procedure and automate the process. The system also takes less time and produces highly accurate results in terms of human observation.