how to do pcolor plots from 3d data?

Hi all,

I have a data set of the type

x=100;
y=100;
z=100;
r = randi([1 5],x,y,z);

I would like to plot a contourplot from pcolor or contourf (2D data from r in z=60) which colors only all number 1's (like r(:,:,60)=1) with red color, in z=60. The other numbers are represented by the transparent pixels.

I tested

f  = figure(1);
pcolor(r(:,:,60)==1);
shading interp;
hcb=colorbar;

But does not work. Can anybody help me?

Thank you so much in advance.

Answer 1

[Walter Roberson]

x=100;
y=100;
z=100;
r = randi([1 5],x,y,z);
cmap = [0 0 0; 1 0 0];
mask = double(r(:,:,60) == 1);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);
set(gca, 'XDir', 'normal', 'YDir', 'normal');

figure
h = pcolor(mask);
h.AlphaData = mask;
colormap(cmap)
set(gca, 'XDir', 'normal', 'YDir', 'normal');

If you look carefully, you will notice a lot of differences. Remember that pcolor() interpolates faces, pcolor() does not produce images -- pcolor() produces surfaces ... viewed from above. pcolor() is literally surf() followed by view(2) and surf() treats the input data as vertex colors and interpolates the color of faces from the four corners.

I would suggest to you that it would make more sense for you to use image() or imagesc() than to use pcolor()

[SAC CSA]

Hi @Walter Roberson, thank you very much for the answer. I tried apply the code to the data attached (which have the same structure of r asked above, except by the dimension 50x50x50), but it does not print any number 1 with respect to z=25, from the data attached.

dc = load('savedata2.mat');
dc1 = struct2cell(dc);
r = cat(3,dc1{:});
cmap = [0 0 0; 1 0 0];
mask = double(r(:,:,25) == 1);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);
set(gca, 'XDir', 'normal', 'YDir', 'normal');

The output shown is

However, there are elements 1's in r(:,:,25), that should be printed. I thought the problem might be in the 'mask' variable. You can help me edit the command to obtains pixel red to relative to 1's?

[Voss]

There are no 1's in r

dc = load('savedata2.mat');
dc1 = struct2cell(dc);
r = cat(3,dc1{:});
find(r == 1)

ans =

  0×1 empty double column vector

ur = unique(r);
[~, idx] = min(abs(ur - 1));
ur(idx)

ans = 1.0000

ur(idx) - 1

ans = 4.4176e-07

[SAC CSA]

Hi @Benjamin thank you for the comment. When showing r(:,:,25) it is possible to see 1's yes. Strangely these 1's are not shown when using find or mask = double(r(:,:,25) == 1); See 1's in the screenshot taken when MatLab to show r(:,:,25).

[Voss]

Those values are close to 1 but not exactly 1. For instance, the element of r that is closest to 1 is actually approximately 1.00000044176, as my comment shows (thanks to @Walter Roberson for adding that point).

[Voss]

Anyway, I thought the premise of the question was that the data only contained integers between 1 and 5, in which case finding elements exactly equal to 1 is all fine and good. This data set obviously does not conform to that premise, so it's not clear what you want the code to do in this case... maybe round each element to the nearest integer first?

[SAC CSA]

@Benjamin, thank you again. Now it's clear to me that the matrix elements are close to 1. In fact, the original question was about painting 1. But since the topic of painting a value close to 1 is related, I thought we could cover it here. I will edit the original question. So I'd like to paint the elemnts of r that is closed to 1 of red pixel. How to do it?

[Voss]

If you want to color the elements that are closer to 1 than they are to any other integer, you can just round the data:

dc = load('savedata2.mat');
dc1 = struct2cell(dc);
r = cat(3,dc1{:});
cmap = [0 0 0; 1 0 0];
mask = double(round(r(:,:,25)) == 1);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

If, however, you want to color the elements whose difference from 1 is less than or equal to some threshold, e.g., within 0.5 or 0.001 of 1, you can do that too:

threshold = 0.5;
mask = double(abs(r(:,:,25)-1) <= threshold);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

threshold = 0.1;
mask = double(abs(r(:,:,25)-1) <= threshold);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

threshold = 0.01;
mask = double(abs(r(:,:,25)-1) <= threshold);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

threshold = 0.001;
mask = double(abs(r(:,:,25)-1) <= threshold);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

threshold = 0.0001;
mask = double(abs(r(:,:,25)-1) <= threshold);
imagesc(mask, 'AlphaData', mask);
colormap(cmap);

[SAC CSA]

@Benjamin, is perfect. Thank you so much for your kindness.

Answer 2

[KSSV]

This is a 3D data, read about slice.