CODEWITHSUNDEEP
matlabplotmatlab-figurecolormap
smyslov
smyslov

Reputation: 1275

How to create a colormap based on value?

I have two vectors as follows: 

x = 0:5:50;
sir_dB = [50 20 10 5 2 0 -5 -10 -20 -20 -20]

Where x denotes the distance on the x-axis and sir_dB the SNR. For this, I need to generate a color map for a grid of 50 x 60m something similar to this:

enter image description here

based on the value of sir_dB.

I tried the following:

sir_dB = [50 20 10 5 2 0 -5 -10 -20 -20 -20];
xrange = 0:50;
yrange = -30:30;
% create candidate set
[X, Y] = ndgrid(xrange, yrange); % grid of points with a spacing of 1. 
candidate_set = [X(:), Y(:)];
test_pt = [0 30];
radius = 5;
% find which of these are within the radius of selected point:
idx = rangesearch(candidate_set, test_pt, radius ); 
neighborhood = candidate_set(idx{1}, :);

Once I have the neighbors at a radius of 5m, I need to color that part of the grid based on the sir_dB value for a corresponding x value.

I need to have the plot in such a way that for all values of sir_dB greater than 15, the grid should be colored green, yellow for y greater than 0 and red for y greater than -20.

Could someone provide me inputs of how to do this best?

Upvotes: 1

Views: 1366

Answers (2)

EBH
EBH

Reputation: 10450

Here is another suggestion, to use imagesc for that. I nothed the changes in the code below with % ->:

x = 0:5:50;
sir_dB = [50 20 10 5 2 0 -5 -10 -20 -20 -20];
xrange = 0:50;
yrange = -30:30;
% create candidate set
[X, Y] = ndgrid(xrange, yrange); % grid of points with a spacing of 1. 
% -> create a map for plotting
Signal_map = nan(size(Y));
candidate_set = [X(:), Y(:)];
test_pt = [10 20];
radius = 35;
% find which of these are within the radius of selected point:
idx = rangesearch(candidate_set,test_pt,radius); 
neighborhood = candidate_set(idx{1}, :);
% -> calculate the distance form the test point:
D = pdist2(test_pt,neighborhood);
% -> convert the values to SNR color:
x_level = sum(x<D.',2);
x_level(x_level==0)=1;
ColorCode = sir_dB(x_level);
% -> apply the values to the map:
Signal_map(idx{1}) = ColorCode;
% -> plot the map:
imagesc(xrange,yrange,rot90(Signal_map,2))
axis xy
% -> apply custom color map for g-y-r:
cmap = [1 1 1 % white
        1 0 0 % red
        1 1 0 % yellow
        0 1 0];% green
colormap(repelem(cmap,[1 20 15 35],1))
c = colorbar;
% -> scale the colorbar axis:
caxis([-21 50]);
c.Limits = [-20 50];
c.Label.String = 'SNR';

The result:

enter image description here

Upvotes: 0

jodag
jodag

Reputation: 22294

Im not sure exactly what you want, but this should get you started with contourf. I increased the granularity of xrange and yrange to make the radius more smooth but you can change it back if you want.

x = 0:5:50;
sir_dB = [50 20 10 5 2 0 -5 -10 -20 -20 -20];
xrange = 0:0.1:50;
yrange = -30:0.1:30;
% create candidate set
[X, Y] = ndgrid(xrange, yrange); % grid of points with a spacing of 1.
candidate_set = [X(:), Y(:)];

test_pt = [0 30];
r = sqrt((test_pt(1)-X(:)).^2 + (test_pt(2)-Y(:)).^2);
idx = r>5;
snr = nan(size(X));
snr(idx) = interp1(x,sir_dB,X(idx),'linear');

% Some red, yellow, green colors
cmap = [0.8500    0.3250    0.0980;
        0.9290    0.6940    0.1250;
        0         0.7470    0.1245];

figure();
colormap(cmap);
contourf(X,Y,snr,[-20,0,15],'LineStyle','none');

Plotting the the contour plot alongside the original sir_dB we see that it lines up (assuming you want linear interpolation). If you don't want linear interpolation use 'prev' or 'next' for the interp1 method.

figure();
colormap(cmap);
subplot(2,1,1);
contourf(X,Y,snr,[-20,0,15],'LineStyle','none');
subplot(2,1,2);
plot([0,50],[-20,-20],'-r',[0,50],[0,0],'-y',[0,50],[15,15],'-g',x,sir_dB);

enter image description here

Upvotes: 1

Related Questions