Currently I work on the new approach to multigrids methods for Poisson equation resolving. It continues my recent work, described in paper Accurate RT Disparity Estimation with Variational Metod.
A pretty interesting result, that I would like to publish here is the residual r(x,y) = f(x,y) + L[u(x,y)] scaled by factor of 1024. The problem is formulated as follows: constant boundary conditions (u(x,y) = 128 on boundaries) and two knots inside the area - a sink (left) and a source (right) (f(width/4, height/2) = -128 and f(3*width/4, height/2) = 128).
Both residual clouds look like a random Gaussian distributions, but nevertheless there are a pretty large difference in distributions for the positive and negative peculiarities. (The absolute value of the residual is shown).
Works In Progress
1 post • Page 1 of 1
Users browsing this forum: No registered users and 1 guest