基于前向和后向扩散的图像增强算法的改进

Improvement of Image Enhancement based on FAB Diffusion

图像去噪和增强是图像处理和计算机视觉领域中的基本问题,而偏微分方程已经广泛应用于模糊图像的复原。针对P-M方法和原FAB方法的不足,通过区分图像的平坦区和边界区,综合这两种方法得出了新的扩散系数方程,并通过有限差分法将对应的偏微分方程离散化后得到了它的数值解。这种改进的各向异性的扩散方法,在平滑图像的同时能够保持和增强边界,对实际图像的滤波结果表明了该算法是有效的。

Image denoising and enhancement are fundamental problems in the field of image processing and computer vision, and Partial Differential Equations （PDE） have been wildly used to restore noisy blurred images. Aimed at the defects of both the P-M diffusion and the FAB diffusion, the two algorithms are combined to produce a new equation of the diffusion coefficient, by distinguishing between the homogeneous parts and the edges. An efficient, fast numerical solution to the new PDE＇ s is found by careful selection of finite differences which propose a discrete algorithm. Images are smoothed while still retaining and enhancing edges by the improved anisotropic diffusion method, and the corresponding filter implementation on a real image demonstrates the efficiency of the scheme.