How do I plot the data as patches rather than vertices in a grid or a mesh?

I have been trying to visualize a set of stochastic data like the matrix below:

     1     1     1     2     2     3     3;

     1     1     1     2     2     3     3;

     1     1     1     1     2     2     3;

     1     1     1     1     2     2     3;

     1     1     1     1     2     2     3;

I want to create something similar to a chess board, with 1 being blue, 2 being yellow and 3 being red (see the attached figure).

I have tried contourf and surf, but they all put data on the vertices of the lattices, so if the shading function is faceted, it will give

And if the shading is interp, it will be

The boundaries between ones, twos and threes is an artificial transition at the blank area where no data exist.

I then tried adding an additional row and column, or, as Youssef KHMOU suggested, using the kron function. Both ways gave the right lattice pattern, but then I had to manually adjust the position of the ticks to produce the first graph, which is what I need.

Is there a better way to do this?

Answer 1

[Star Strider]

Use the pcolor function.

A =  1  1  1  1  1  1  1;
     1  1  1  1  1  1  1;
     1  2  2  1  1  1  2;
     2  2  2  2  3  3  2;
     2  2  3  3  3  3  3];

figure(1)
pcolor(A)
axis equal
colormap(gray)

[Kun]

I tried you way, but the result were the same. Guess I did not make my question clear.I will rephrase it.

[Star Strider]

This may be what you want:

A = [1  1  1  1  1  1  1;
     1  1  1  1  1  1  1;
     1  2  2  1  1  1  2;
     2  2  2  2  3  3  2;
     2  2  3  3  3  3  3];

x = [0:6]+0.5;
y = [0:4]+0.5;

figure(1)
image(x, y, A)
axis equal tight
colormap(gray(3))
grid on
h = get(gca)
set(gca, 'YTick', 0:5, 'GridLineStyle','-')

produces:

Answer 2

[Youssef Khmou]

Tensor product can be efficient in representing each number with different color, but try this way i think it is convenient to your description :

 M=ones(4,1)*(1:4)
 surface(M)

[Kun]

I tried your way, but unfortunately it did not work.Maybe I misunderstood you meaning. What dose multiplying by ones() do to my data?

[Youssef Khmou]

it was only a way to create a matrix similar to that you posted, so surface is not working?

[Kun]

surface gave a similar result as surf did.

Answer 3

[Kun]

I just tried adding an additional row and column to the original matrix and then using the surface function. It gave the correct chess board pattern I need, but the ticks on x,y-axis were still assigned to the vertices, so I had to shift them to match the location of the patches. But this is tedious if the size of the data matrix is huge. Does anyone have better solutions?

[Youssef Khmou]

try :

M=ones(4,1)*(1:4);
B=kron(M,ones(3));
figure; surface(B); 
% shading interp

[Kun]

Yeah, it worked. But I still need to adjust the position of the ticks on both axes. Perhaps the only way to get the graph I want is to do it in two steps.