Image processing logic implementation

Hi all,

I'm trying to implement this algo from a journal, and I'm stuck at line 13. I've completed all the above steps, but not able to think of logic for line 13.

It says that use the matrix extracted from noise-free pixels which have a distance=1 i.e w(d(1)) matrix. So, in the first iteration when k=1, i'll use matrix corresponding to this matrix and continue with next steps. Next iteration k=2, I'll use matrix corresponding to a distance=2.

My question is in every iteration for k=1,2,3,4 how can I use a new matrix for each new iteration.

The procedure has shown in image below.

I understand it might be an easy algo, but I'm facing difficulty in logic implementation?

Please advise.