图像修复RBF模型的Kalman改进算法

Improved Image Restoration RBF Model Based on the Kalman Filter

利用Kalman滤波器训练图像修复RBF网络模型的权值与基函数的中心,由于利用滤波器测量噪声的协方差阵实现整个模型的图像降噪功能,新算法边缘修复能力较强.本文通过仿真,检验了新模型的降噪能力;结果证明,新算法较一般的维纳滤波器,均值滤波降噪能力明显提高.

This paper readopts a new kind of Kalman filter to train the Radial Basis Function (RBF) neural networks weight and the centers of the base functions, through setting the measurement noise s covariance matrix, the new algorithm realize noise-filtering capacity of the model, the remarkable character of the new algorithm is its powerful boundary restoration capacity;Through the simulation, this paper compare the new algorithm and the wiener and mean filter algorithm, the result shows that the new algorithm improve the models performance perfectly in the increment of the signal noise ratio aspects.