Edge Detection With Diffusion Filter

Edge detection in image processing and plays an important role, the merits of the algorithm will directly affect the performance of the system under study. The traditional edge detection algorithm has been great development, but there are still many problems, especially in edge detection accuracy and noise immunity, coordination, how to reconcile the contradiction between the two and seek algorithm is simple and easy to implement edge detection algorithm, is a very meaningful topic to be studied. In fact, the margin of the target object, texture and even the noise is likely to be meaningful edge information, so for different edge features difficult to find a universal adaptive edge detection algorithm. Various existing edge detection algorithm is the edge of the image for certain characteristics designed to detect effects in the same time there are also their own limitations and inadequacies, the digital image edge detection algorithm needs to be further improvement and development. In view of this, in the following aspects of this thesis work undertaken:The theory of diffusion filtering in image edge detection, in the diffusion filtering of images, the use of bi-proliferation framework, that is the image of the "flat" area we use diffusion to smooth the noise, the edge of the image we inhibited proliferation, through control of the diffusion coefficient can be increased to reverse spread of the edge effect. Mathematical framework of the improved and optimized. A new diffusion filter edge detection algorithm, and apply the simulation results show the effectiveness of the algorithm and strong adaptability. This method not only can remove noise, to retain details of the image features; the same time that the method can also repair images, an enhanced effect on the image...