作者: J. Zerubia , R. Chellappa
DOI: 10.1109/72.238324
关键词:
摘要: The authors consider the problem of edge detection and image estimation in nonstationary images corrupted by additive Gaussian noise. noise-free is represented using compound Gauss-Markov random field developed F.C. Jeng J.W. Woods (1990), posed as a maximum posteriori problem. Since probability function nonconvex, computationally intensive stochastic relaxation algorithms are normally required. A deterministic method based on mean annealing with (CGMRF) model proposed. present set iterative equations for values intensity both horizontal vertical line processes or without taking into account some interaction between them. relationship this technique two other methods considered. Edge results several noisy included. >