Flaw detection in granites by image analysis

Image Processing has a vital role in quality analysis of objects and presumably is one of the best ways of non destructive techniques. This paper aims at detecting flaws if any in the granite blocks by analysis of their digital images. Two common defects, Crack and Mica contamination is studied. The unique characteristics of the defects are studied. To detect the mica contamination k-means segmentation algorithm is used. For crack detection, linear edge detection algorithm is applied and further processing is done to discriminate edges of cracks. Thus the algorithm categorizes the granite as defect less or defected granite. The efficiency of the algorithm has been justified subjectively with the sample images and their outputs and objectively with the parameters of the Confusion Matrix.