摘要
从观测模型和先验模型入手,深入分析了在图像去噪过程中如何在边缘保护和去除噪声之间寻找平衡,针对TV先验模型在边缘保护和纹理保护中的不足,提出了新的基于最大后验概率的la范数去噪方法,该方法更好地保护了图像的边缘,同时使得在平滑区域产生的阶梯效应减少.数值实验进一步验证了该方法的正确性和有效性.
In the paper,from the observation model and a priori model,the problem of image denoising is discussed,and how to find a balance between edge protection and image denoising is analyzed in depth. In addition,aiming at the advantages and disadvantages of TV prior model on the edge protection and texture protect,the new denoising method named la norm based on Maximum A Posteriori Probability（MAP） is put forward. This method provides better protection of the edge of the image. Furthermore,the staircase effect in smooth regions is reduced. Numerical simulations are further provided to verify the feasibility and effectiveness of the theoretical results.
作者
宁娣
岳晨曦
Ning Di Yue Chenxi(College of Mathematics and Statistics, South-Central University for Nationalities , Wuhan 430074 , Chin)
出处
《中南民族大学学报（自然科学版）》
CAS
北大核心
2017年第1期128-131,共4页
Journal of South-Central University for Nationalities：Natural Science Edition
基金
全国统计科学研究项目(2014LY020)
