Research and Application of Metallographical Image Edge Detection Based on Mathematical Morphology

Aimed at the characteristic of metallographical image, a specific image, and based on the multi-structural element in the mathematical morphology, a new method of edge detection is proposed in this paper to deal properly with the double- boundary and the inaccuracy of the boundary localization in the image in the edge detection carried by the traditional edge detection method, such as Roberts, Sobel, Prewitt, Log, Canny and so on. The design of the multi-structural element and the construction of the suitable morphology operator are emphasized in this paper. The experimental results indicated that this method is very effective, and is better than the traditional edge detection method.