自适应Volterra混合噪声图像滤波

采用更接近实际的随机值脉冲噪声对混合噪声图像中的脉冲噪声部分进行描述，基于自适应Volterra滤波方法，提出了一种能够有效去除图像混合噪声的LMSV滤波算法．在此基础上，针对LMS算法收敛速度慢，步长因子不易确定的特点，提出了一种改进的自适应Volterra图像滤波算法（NLMSV算法）．理论分析和计算机仿真实验表明，相比于传统的图像去噪方法，本文方法在去除图像混合噪声和保留图像边缘细节方面的性能要明显优于其他算法，并且图像受污染率越高，该方法表现出的性能越好．

Adaptive Volterra Image Filtering for Mixed Noise Removal

The random-valued impulse noise is defined to describe the impulse noise part of the mixed noise in image,which is more suitable for realistic noise conditions. Based on adaptive Volterra filtering method, the LMSV image filter algorithm is proposed for the removal of mixed noise effectively. For the characteristics of the slow convergence of LMS algorithm and the difficulty of step factor determination, a modified adaptive Voherra image filtering algorithm （ NLMSV algorithm） is put forward. Theoretical analysis and computer simulation re- sults show that the new method performs superior to traditional image denoising methods in removing mixed noise as well as preserving edge details, and the higher the image contaminated, the better the new method performs.