摘要
研究图像平滑处理去噪问题。针对传统图像平滑算法在去除噪声的同时会破坏边缘、纹理等不能保持原图图像特征。为解决此问题,提出了一种基于PDE's各项异性的扩散Perona&Malik算法。通过引入高斯平滑算子,利用P&M模型的可调扩散系数来改善图像平滑效果,在抑制噪声的同时能够保持这些特征的特点。通过仿真分析表明,所提出的算法对孤立的噪声点平滑效果明显,同时也不会过多地影响原来图像的特征。
Research for image smoothing processing denoising question. The traditional image smoothing algorithms destroy the edge in eliminating noises, texture and other image characteristics can not be maintained as the original image. To solve this problem, a heterosexual spread of Perona & Malik algorithm is proposed based on PDE's. By introducing Gaussian smoothing operator, P & M model uses the adjustable diffusion coefficient to improve the smoothing effect of the algorithm and reduces noise without destroy the characteristics of these features. The simulation analysis shows that the proposed algorithm has obvious effect to isolated noise points, and not much affects the original image feature.
出处
《计算机仿真》
CSCD
北大核心
2011年第8期256-258,279,共4页
Computer Simulation
关键词
各向异性
梯度阈值
图像平滑
Anisotropy
Gradient threshold
Image smoothing