Is there a way to extract data from surface graphics in the curvefitting toolbox?

Hello to all

I'm doing some interpolation problems using curve fitting toolbox.

I would like to know if there is any way to extract the x, y z of the surface graphic that appear in the curve fitting window?

Thank you

Pedro

[Walter Roberson]

There might be. But it would be easier to save the fit to workspace and use the fit together with x y coordinates to project z coordinates .

[Pedro Pinto]

The question is that we generate a surface graph with the curve fitting toolbox and want to use the same x,y,z data to create a heat map on another image software. The dificulty is that we cannot get the x,y,z variables of the interpolated graphic to use them in the next phase. If anyone has some ideas we thank in advance

Pedro

[Walter Roberson]

Tomorrow I will do some testing on how to extract the exact information . It will not be easy as it involves undocumented graphics containers if I recall correctly .

Answer 1

[Walter Roberson]

Experimenting, I find:

s = findall(groot,'tag','curvefit.gui.FunctionSurface');
x = s.XData;
y = s.YData;
z = s.ZData;

Answer 2

[Pedro Pinto]

Dear Walter

I was trying and it give an error

The code, generated by curvefitting toolbox is this one:

function [fitresult, gof] = createFit1(x, y, Pontos)
%CREATEFIT1(X,Y,PONTOS)
%  Create a fit.
%
%  Data for 'untitled fit 1' fit:
%      X Input : x
%      Y Input : y
%      Z Output: Pontos
%  Output:
%      fitresult : a fit object representing the fit.
%      gof : structure with goodness-of fit info.
%
%  See also FIT, CFIT, SFIT.

%  Auto-generated by MATLAB on 17-Feb-2019 16:25:20

%% Fit: 'untitled fit 1'.
[xData, yData, zData] = prepareSurfaceData( x, y, Pontos );

% Set up fittype and options.
ft = 'thinplateinterp';

% Fit model to data.
[fitresult, gof] = fit( [xData, yData], zData, ft, 'Normalize', 'on' );

% Plot fit with data.
figure( 'Name', 'untitled fit 1' );
h = plot( fitresult, [xData, yData], zData );
legend( h, 'untitled fit 1', 'Pontos vs. x, y', 'Location', 'NorthEast' );
% Label axes
xlabel x
ylabel y
zlabel Pontos
grid on

where x, y and Pontos are the vectors of x,y,and z containing the data

[Walter Roberson]

You did not indicate what error message you get ?

Example data would also help.

I just tested again, and the code I suggested earlier still appears to work. However, since your code is calling plot() itself, there is a small shortcut: instead of s = findall(etc), make it s = h(1); where h is the value assigned by the plot() call.

[Pedro Pinto]

I tryed and now there are no more erros and i get a variable called ans with 1x1 sfit as value.

However i cannot get the values of the vectors in the full space.

So let me get the values

i have a 6 x 6 table of x,y data

x=1,2,3,4,5,6

y=1,4,2,7,4, 9

then for each x,y pair (ex, 1 x 1, 2 x 4, etc), i have the following z data:

Pontos: 43, 65, 7, 45, 98, 34

When i create the interpolation graphic in the curvefitt toolbox i get a surface continues graphic with x between 0 and 6, y between 0 and 10 and z between 0 and 100.

These are the values that i need to have in 3 new variables x1, y1 and z1 so i can pass them to excel and enter in the other software. It will be a continue surface x,y,z combinations based on the thin plate spline interpolation

For sure this is a simple question, but as this is my first approach to matlab world it's being difficult to get the values

[Walter Roberson]

Please post your data for testing.

Answer 3

[Pedro Pinto]

Here follows

[Walter Roberson]

With that data, the code I posted works for me, both versions -- doing the findall() after the createFit1 routine has returned, or taking h(1) of the h = plot() handle in the createFit1 routine.

The x, y, and z values that are extracted are all 51 x 49, and you can surf(x, y, z) to reconstruct the plotted surface.