Plotting a 2D Contour from data (A variable with its XY coordinates)

Hello!

I need to plot a 2D Contour form data. I have a variable called "Intensity" with 224x1 values and I have a variable called "XYAxis" with the corresponding coordinates of this Intensity (224 points (X,Y)).

What i want to do is to plot both variables, to have a contour2D with the distribution of the Intensity among the points analized (each value of Intensity corresponds to one point(X,Y).

I have tried:

Contour (XYAxis,Intensity);

But it only displays an empty plot and the axis are not those i desire (i want the contour to be like scattered).

Any idea? Maybe TriScatteredInterp()?? but i need the contour to be 2D

Thanks in advanced

[Sean de Wolski]

Could you provide a small set of sample data?

[Peter]

Sure,

For example, I have one variable called " XYAxis " (224x2 double) this way (1º column is X and 2º Column is Y):

0 0
0 5
5 10
10 20
20 30
... until 224 points

And I have another variable called " Intensity " like this:

0.5
0.8
0.9
0.10
0.11
...
until 224 values

So each value of "Intensity" corresponds with one value of "XYAxis" And i want to plot it with a contour 2d. Thanks

Answer 1

[Sean de Wolski]

% Synthetic Data  
xy = rand(20,2);
intensity = rand(20,1);

% Interpolate to grid
interpolant = scatteredInterpolant(xy(:,1),xy(:,2),intensity);

% Grid
[xx,yy] = meshgrid(linspace(0,1,10));  % replace, 0 1, 10 with range of your values

% Interpolate
intensity_interp = interpolant(xx,yy);

contourf(xx,yy,intensity_interp)

You'll have to interpolate it to a grid and then call contour.

[Peter]

Thank for the explanation and the code, but when i run it, it displays this error message:

??? Undefined function or method 'scatteredInterpolant' for input arguments of type
'double'.

[Peter]

Both variables "XYAxis" and "Intensity" in my case are double

[Sean de Wolski]

You must be using an old release. TriScatteredInterp was the predecessor to scatteredInterpolant and griddata the predecessor to TriScatteredInterp.

For this example, I believe scatteredInterpolant can be directly substituted for TriScatteredInterp. Of course, if your company or university if current on maintenance, you can upgrade to the newest release for no additional cost.

[Peter]

Thank you very much for that info. I tried with TriScatteredInterp and it worked.

However, i would like to know if by defining linespace(....) i am fixing the size of my axis...because my axis are not a square (i mean, with linespace my plot displays same lenth for x axis and same lenght for y and my data is not this way).

I want the mesh to be just the points of my data (only my X and Y coordinates) no others. (Like scattered).

Do you know how could i do it?

Thanks!

[Sean de Wolski]

xv = linspace(min(xy(:,1)),max(xy(:,1)),20);
yv = linspace(min(xy(:,2)),max(xy(:,2)),20)
[xx,yy] = meshgrid(xv,yv)

The grid does not need to be square.

You could also use

ux = unique(xy(:,1))
uy = unique(xy(:,2))

etc.

[Peter]

Thank you very much, it really worked and I understood your code.

Could you please tell me, if you know how to save the contour plot into an image , for example jpg?. Wich image format would your recommend?

Thanks

[Sean de Wolski]

For single frame images, I usually use PNG since it's really common, uses lossless compression, and supports transparency.

You can use print/saveas/hgsave etc.

saveas(gcf,'test.png')

[Peter]

I mean "automatically", with the names of the axis on it and a legend

[Sean de Wolski]

xlabel('whatever x is')
ylabel('whatever y is')
filename = ['file' num2str(1) '.png'] % replace 1 with whatever you want automatically created.
saveas(gcf,filename)