摘要
关于图像获取过程中的优化问题,为了克服传统非线性扩散图像去噪模型的不足,有效去除图像噪声,提出了引入两项构造的控制函数。利用局部背景方差信息构造ε方向扩散控制函数以更好区分边缘和噪声,并对边缘区域内,边缘区域间及不同背景下的噪声采取各自适合的扩散策略;采用局部背景亮度信息构造η方向扩散控制函数以加速区域内平滑,消除"块效应",约束去噪图像更逼近原图。理论分析和实验结果均表明,算法模型既能较好保持图像的边缘结构,又能充分抑制噪声,并对视觉效果和峰值信噪比值上都具有明显优势。
In order to overcome the disadvantages of traditional nonlinear-diffusion-based image denoising methods,two control functions are constructively proposed to denoise image effectively.Using the information of local background variance,a ε-direction diffusion controlled function is constructed to have a better distinction between edge and noise,and adopt suitable diffusion strategy to the noise in inter-edge-region and intra-edge-region with different backgrounds respectively.Using the information of local background brightness,a η-direction diffusion controlled function is constructed to accelerate the inter-edge-region smoothing and eliminate ＂blocking effect＂,so that the denoising picture is more approaching to the original one.Theoretical analysis and experimental results demonstrate that the new algorithm model can both restrain noise better and hold edge structure fully,as well as eliminate ＂blocking effect＂ efficiently,and there is superiority in the visual effects and peak signal noise ratio.
出处
《计算机仿真》
CSCD
北大核心
2010年第12期237-240,共4页
Computer Simulation
基金
重庆市科委基金项目(CST2005BB0061)
重庆市科委基金项目(KJ070514)
关键词
图像去噪
非线性扩散
局部背景方差
局部背景亮度
视觉系统
Image denoising
Nonlinear diffusion
Local background variance
Local background brightness
Visual system
