Research On Medical Image Edge Detection Based On Mathematical Morphology

Medical image plays a vital role in the clinical diagnosis and remedy. The doctor can't judge the result easily because of its fuzzy edge and the noise interference. As a key part in the digital image processing and computer vision, edge detection is the basement for the image analysis and recognition, and can be used to improve the quality of the medical image. So many people pay close attention to this aspect. As a new subject in the image processing, medical image edge detection based on mathematical morphology is of great practical significance and has important prospect. It is always a hot topic in the field of automatic target recognition.As far as I'm concerned, the main research of this paper is to detect the edge of medical image based on mathematical morphology. The basic theory of edge detection and mathematical morphology are introduced in the paper at first. The classic edge detection algorithms are analyzed particularly by the simulation. Considering the characteristics of medical image, an improved median filtering algorithm based on adaptive threshold is proposed in order to overcome the noise in the medical image. This method can remain the image details and reduce the noise effectively. Combining the characteristics of the traditional morphological edge detection algorithm and features of medical image, a general morphological blur-minimization edge detection algorithm based on multi-scale and full-direction structure elements is put forward for medical image on the fundament of previous research. This improved edge detection algorithm which is based on the general morphological edge detectors can make use of multi-scale and full-direction structure elements to extract the edge information, then according to the suitable weights combine the result to make up another edge image which can gain the final edge through blur-minimization edge detection. The proposed algorithm is applied to detect the edge of medical image as a simulation in the environment of MATLAB 7.0. The experimental results and the quantitative date demonstrate that the proposed method can not only eliminate the noise in the medical image efficiently, but also extract a suitable edge. So it is practical in the medical image processing. It has a good performance in helping doctor to obtain a better diagnosis as well.