A novel model using Kalman filtering for image restoration

摘要

A novel Kalman filtering model and algorithm is proposed in the paper. We give the state equation from the Euler-Lagrange equation of the total variation model. The nonlinear partial differential equation is added with a gaussian white noise and discretized by finite difference method. Then the parameters of the state equation can be derived from the discrete equation directly. We also discuss some numerical experiments which prove our proposed model and algorithm to be more efficient.