Program the colormap of a graph using different colors for different ranges

Dear all,

I have a matrix containing NaNs and numbers:

    x=[NaN  1.2  68.65  155.8  2.41  1.81  .2;
4.41  NaN  45.19  127.2  6.26  72.73  1.6;
35.39  28.59  NaN  37.36  94.5  4.87  48.62;
78.4  93.64  4.64  NaN  181.99  17.14  128.6;
8.35  29.8  9.23  151.1  NaN  128.6  1.7;
53.43  55.42  8.71  14.93  117.62  NaN  74.61;
.7  1.12  21.95  56.4  13.43  35.88  NaN];

Now, I would like to create an image with a colorbar where the colors change according to specific ranges:

• NaN -> [255 255 255]
• 0 < x < 0.001 -> [197 215 232]
• 0.001<x<0.01 -> [132 170 206]
• 0.01<x<0.1 -> [89 141 202]
• 0.1<x<3.8 -> [0 0 255]
• x>3.8 -> [255 0 0]

What I did is this code:

    Results_Sorted=sort(x(:)); % sort data

    map=zeros(size(Results_Sorted,1),3,'double');% create the colormap

    %looking for indexes of value ranges
    pos_0p001_min= sum(Results_Sorted<=0.001);

    pos_0p01_min=sum(Results_Sorted<=0.01);

    pos_0p1_min=sum(Results_Sorted<=0.1);

    pos_1_min=sum(Results_Sorted<=1);

    pos_3p8_min=sum(Results_Sorted<=3.8);

    % fill the map with my colors

    for i=1:pos_0p001_min

    map(i,:,:)=[255 255 255];

    end%for

    for i=pos_0p001_min+1:pos_0p01_min+1
    map(i,:,:)=[197 215 232];

    end%for
    for i=pos_0p01_min+1:pos_0p1_min+1
    map(i,:,:)=[132 170 206];
end%for
for i=pos_0p1_min+1:pos_1_min+1
    map(i,:,:)=[89 141 202];
end%for
for i=pos_1_min+1:pos_3p8_min+1
    map(i,:,:)=[0 0 255];
end%for
for i=pos_3p8_min+1:size(map,1)
    map(i,:,:)=[255 0 0];
end%for
map=map/255;

    imagesc(Result);
    colorbar;
    colormap(map);

But this is not exactly what I need... Could you help me to understand where is my error? Best regards. Sébastien

Answer 1

[Guillaume]

My suggestion would be to inverse the problem: Have a simple colour map with just your six entries and discretise your x so that it's only got six different values. You can adjust the colourbar ticks so that it displays the real x values instead of the discretised values:

map = [255 255 255
       197 215 232
       132 170 206
        89 141 202
         0   0 255
       255   0   0] / 255;
fakex = discretize(x, [0 0.001 0.01 0.1 3.8 Inf]) + 1; %+1 since NaN is the first colour
fakex(isnan(x)) = 1;
imagesc(fakex);  %note that your example x is not very good since it only includes the last 2 intervals
colormap(map); 
hbar = colorbar;
%option 1: label start and end of each colour box
hbar.Ticks = linspace(1, 6, 7);
hbar.TickLabels = {'NaN', '0', '0.0001', '0.01', '0.1', '3.8', 'Inf'}
%option 2: label middle of each colour box
hbar.Ticks = conv(linspace(1, 6, 7), [0.5 0.5], 'valid');
hbar.TickLabels = {'NaN', '0 - 0.0001', '0.0001 - 0.01' '0.01 - 0.1', '0.1 3.8', '3.8 - Inf'}