摘要
针对乘性噪声去噪算法研究,乘性噪声多出现在合成孔径雷达、超声波和激光等相干图像系统中,与标准高斯加性噪声不同,乘性噪声符合瑞利和伽马分布函数。文中通过:(1)取对数把乘性噪声模型转变成相加形模型;(2)改进正则化项成为自适应扩散模型;(3)将数学模型应用于图像处理的实践当中。在解决了阶梯效应的同时保持了图像的边缘。
Multiplicative noise, unlike the standard Gaussian additive noise, often appears in synthetic aper- ture radar and sonar and laser imaging. The muhiplicative noise agrees with Rayleigh and Gamma functions. This paper converts the multiplicative model into an additive one by taking logarithms, changes the standard TV regularizer into an adaptive fidelity term diffusion model and applies mathematical models into multiplicative imaging denoising. The model can not only remove noise but also enhance edges.
出处
《电子科技》
2011年第1期1-3,8,共4页
Electronic Science and Technology
基金
国家自然科学基金资助项目(60872138)
关键词
乘性噪声
图像恢复
总变分
自适应正则化项
multiplicative noise
image restoration
total variation
adaptive fidelity term