Automatic Tile Inspection

Abstract

In tile industry, quality of a tile should be maintained. But manual inspection is time consuming and less efficient. This report will help to automate inspection of tiles. For detecting edges of an object from an image various edge detection techniques like sobel, prewitt, canny, laplacian, Marr-hildreth can be used. Hough transformation and Contour techiniques are used for detecting exact tile from an image. For detecting defects some operations or set of operations like morphological operations, segmentation can be used. Canny edge detection algorithm is used in the proposed system as well as for better detection of objects, some preprocessing steps are required which are also mentioned. Techniques for detecting surface level defects are also proposed. Thus proposed system will help the inspection of tiles to be qualitative and less time consuming.