复数域非线性扩散滤波在图像处理中的应用

图像处理中去噪和边缘保持是一对矛盾体,去噪滤波经常会失去高频部分的信号,而边缘信息主要以高频信息为主.考虑了一种新的方法——复数域非线性扩散方法,该方法通过结合自由薛定谔方程将传统各向异性扩散方程拓展到复数域,复数域实部信号可以很好地滤波,将复数域扩散项虚部加入shock滤波器,该方法克服了传统的边缘检测方法边缘保持能力差的缺点.试验结果表明:该算法具有良好的抗噪性和边缘细节的保持性,边缘检测效果良好.

Complex Field of Donlinear Diffusion Methods in Image Processing

Denoising and edge preserving is a contradiction in image processing,denoising filter often lose the high frequency part of the signal,However,the edge information based mainly on high-frequency information.This paper introduces a new method-nonlinear complex diffusion method,the method by combining free Schrodinger equation will be the traditional anisotropic diffusion equation is extended to complex number field,complex number domain real signal can be very good filtering,the imaginary part of complex diffusion item add shock filter,this method overcomes the shortcomings of the traditional edge detection methods edge preserving ability difference.The test results show that:the algorithm has good noise immunity and edge retention,edge detection effect is good.