Adaptive cell-centered finite volume method for diffusion equations on a consistent quadtree grid (Q273656)

Image processing problems are frequently modelled by nonlinear partial differential equations. In instances where irregular finite volume grids are employed, one needs a good approximation of the gradient function. This is especially true when a coarsening is done based on quadtree grids. The authors present a construction of a deformed quadtree grid in such a way that one could apply the classical finite volume methods. A new cell-centered finite volume scheme to evaluate solution gradients is presented. This reduces to the problem of solving a system of linear algebraic equations. Unique solvability of such a system is shown. Numerical experiments are performed on the regularized Perona-Malik model where the proposed method is employed.