Based On The Double Structure Of Mathematical Morphology Edge Detection Method

The edge which is widely used in image processing such as feature description, image segmentation, image enhancement, image compression and pattern recognition etc. is one of the most important fundamental feature of an image. It is not so much that we argue the importance of the edge of image. There is a plethora of papers in the subject and are so many scientists who work on this problem and derive a lot of edge detection filters and algorithms that are various degrees of success of different image. Conjoint pixels, which gray-level change greatly are considered image edge .In mathematics, this distributing change is often depicted by gradient. In this paper, we research several traditional and new type algorithms, some of which base on differential, and analyze their merit and disadvantage.Detecting image edge by mathematics morphology is the main goal of this paper .So we introduce origin of mathematics morphology from binary morphology to gray morphology and extensively study its different operators and quality .We happen to find that some mathematics morphological operators have ability of resisting noise, at the same time, all images have certain noise which influence detecting result except ideal image, so our research has practical value.In the paper, we extend the definition of traditional image edge to morphology edge, derive a series of new morphological operators which are enlightened by semi-increasing and semi-continuation and first time bring forward double-structure element used in morphological processing at one time. These new morphological operators can resist noise and have good merit, which not only have good location but also keep image detail. Experiment proves that these operators can well detect imageedge.