Recognizing a chip on edge of a cutting tool by Hough transform

Conventionally, finished cutting tools are checked by human eyes. Most of the defects are minute and it is needed to be checked with a microscope. Therefore, it is a very labourous work. To lighten human load, the automatic testing system is required. The aim of this research is the development of a recognizing system for a chip on the edge by using image processing. As the preprocessing methods, a color image of the edge is analyzed through grayscaling, binarization, extraction of the outline by Prewitt-filter and Hough transformation. These procedures are developed and realized in the developed testing system to make the edge detection.To ascertain the effect, estimation of the amount of a chip on the edge is made not only the judgement if it has a chip or not. And using neural network, an abnormal chip, a broken edge and a surface defect are recognized.