An Image Edge Detection Algorithm Based on Improved Mathematical Morphology and Wavelet Transform

Abstract

Edge detection of medical images is a very important part of diagnosing the condition. Aiming at the problems that the traditional canny edge detection algorithm is easy to lose edge details and weak noise suppression, we propose a fusion of improved mathematical morphology and wavelet transform edge detection algorithm. Firstly, the image is decomposed to three-level wavelet. With the improved mathematical morphology, edge detection is performed on low frequency sub-images, while hig-h frequency sub-images is denoised by improved filters.Then based on wavelet modulus maximum, Otsu method is used to accomplish threshold edge detection on images. At last, apply a new fusion way to fuse the results of the two methods together.The experimental results show that the improved algorithm has better noise suppression ability for medical image detection, and the detected image edge connection is better.