Smooth data with correlation

Hello all,

I need to create a file containing random numbers (let's say on a 64x64 grid) and then smooth the data to create a correlated function with an arbitrary correlation distance L. I basically need to create something like this :

The random grid is easily created using

random_grid = (b-a).*rand(64,64) + a;

[a,b] being the interval in which the random numbers are generated.

• I have trouble finding a way to smooth the data that creates bumps on the surface that have the desired size;
• If i also could control the maximum local slope that'd be top tier.

Does anyone know a way to do this ? Thanks !

Answer 1

[KALYAN ACHARJYA]

data_image=randi(255,64,64);
% Smooth the image as per your requirements
% or Bluring the image see the fspecial documentation
h=fspecial('average',3);
data_result=imfilter(data_image,h);
image(data_result);
colorMap=jet(256);
colormap(colorMap);
colorbar;

Detail fspecial here

Similar question answer see here (@Image Analyst)

Also check with

h=fspecial('motion',some_value,some_value);

Hope it helps!

Answer 2

[Gregoose]

Thanks for your answer, I'll now have to play with the kernel size/type to fit my needs, as well as the edges !