3D plotting, function approximation

Hello !

I do have some problem with plotting 3d function.

I have 3 column vector (X1,X2,Y) (X1,X2 is the input coordinate Y is the value in (X1,X2) coordinate) Y has the same dimension as X1,X2. I wolud like to plot Y in 3D. I was trying use [A,B]=meshgrid(X1,X2) and after surf(A,B,Y) but Y has wrong dimension to use surf(). Can anybody help, how can I transform Y to have the proper dimension? (The function between (X1,X2) and Y is unknown, I now only the value of Y in the current points)

Thank you!

[Cam Salzberger]

Hello Peter,

Are X1 and X2 evenly spaced? Like:

X1 = [1 2 3 4 1 2 3 4 1 2 3 4];
X2 = [5 5 5 5 6 6 6 6 7 7 7 7];

Because that would make this really convenient. Or are the points totally randomly-placed?

[Peter Kerekes]

Hello Cam !

Accordin to the exercise the points should be randomly selected, but maybe the evenly space is acceptable.

Answer 1

[Cam Salzberger]

Hey Peter,

If the points are random, then generally you would plot them in 3-D as singular points in three dimensions:

plot3(X1, X2, Y, 'ok')

Alternatively, you can depend on Delaunay triangulation to make a surface between the points, and then display that:

tri = delaunay(X1, X2);
trisurf(tri, X1, X2, Y);

If you really want to use "surf", you could create a scatteredInterpolant, make your query points into an evenly-spaced grid, and then display that with surf. That's probably not necessary though. I would guess that delaunay and trisurf would suffice for your needs.

-Cam

[Peter Kerekes]

Hello Cam !

Thanks a lot for your help. I've tried and work well.

Peter,