一个能去噪和保持真实感的各向异性扩散方程

AN ANISOTROPIC DIFFUSION EQUATION THAT CAN REMOVE NOISE AND KEEP NATURALNESS

提出了一个由各向异性扩散方程定义的非线性图像滤波算子．与Ｐｅｒｏｎａ，Ｍａｌｉｋ 和Ｃａｔｔé等人提出的类似算子一样，该算子能够去噪声，而且性能很稳定．另一方面，它也能增强阶跃状强边缘并保持边缘的位置．所不同的是，它还能保持图像中有意义的较强的尖峰和窄边缘．由于这几方面的性能，处理后的自然图像看上去不但清晰度和对比度得到增强，而且有意义的细节也得到保留，很有层次感．该算子特别适用于去除非纹理图像上均匀分布的噪声．实验结果令人满意．

In this paper, a nonlinear image processing operator defined by an anisotropic diffusion equation is presented. Similar to the operators proposed by Perona, Malik and Catté et al. , it can remove noise and performs stably. It can also enhance step like edges and keep the locality of them. What is different is that it is capable of keeping stronger peaks and thin edges. Due to these characteristics, not only the processed noisy images look much more clear and smooth, but also the details are kept, resulting in the naturalness. The new operator is very suitable for removing uniform noise on non textured images. The experimental results are very satisfactory.