# how plot a correct interpolation with surf plot?

조회 수: 15(최근 30일)
Philippe Corner 31 Mar 2020
댓글: Philippe Corner 31 Mar 2020
Im using the code:
figure(1)
X=A_PRE(:,1); Y=A_PRE(:,4); Z=A_PRE(:,5);
[xq,yq]=meshgrid(linspace(min(X),max(X),100),linspace(min(Y),max(Y),100));
zq=griddata(X,Y,Z,xq(:),yq(:),'linear');
[c,h]=contourf(xq,yq,reshape(zq,100,100),'LineStyle','none','LevelStep',100);
hold on
plot(cotas(:,1),cotas(:,2),'k','linewidth',4) %elevation data
hold on
scatter(X,Y,'filled','k')
figure(2)
plot3(Aelec(:,2),Aelec(:,3),Aelec(:,4),'k','linewidth',2);
hold on
scatter3(A_PRE(:,2),A_PRE(:,3),A_PRE(:,4),'filled','k')
hold on
level=A_PRE;
x = level(:,2); y = level(:,3); z = level(:,4); c = level(:,5);
yv = linspace(min(level(:,3)), max(level(:,3)), 100);
zv = linspace(min(level(:,4)), max(level(:,4)), 100);
[Y,Z] = meshgrid(yv, zv);
X = griddata(y,z,x,Y,Z,'cubic');
C = griddata(y,z,c,Y,Z,'cubic');
yA = Aelec(:,3); zA = Aelec(:,4);
in = inpolygon(Y,Z,yA(1:end-1),zA(1:end-1));
Z(in) = nan;
surf(X, Y, Z, 'cdata',C);
for plotting the contour:
using surf function and using X and Y coordinates for its location, but when a use the same view of the 2D contour it seems to be interpolated wrong, any ideas about what could be my mistake? how could a I get a contour plot for figure 2 as similar than figure 1.
As you see, the yellow values seems to be affected with the other data using surf plot, any idea about how to obtain the interpolation plot as similar as fig 1? Thanks in advance

### 채택된 답변

darova 31 Mar 2020
편집: darova 31 Mar 2020
It happens because of different scales of data. Don't know why MATLAB get confused about it (even linear interpolation)
scale and center your data before interpolation
sc = (max(z)-min(z))/(max(y)-min(y));
y = y - mean(y);
x = x - mean(x);
y = y*sc;
x = x*sc;
yv = linspace(min(y), max(y), 50);
zv = linspace(min(z), max(z), 50);
[Y,Z] = meshgrid(yv, zv);
X = griddata(y,z,x,Y,Z,'cubic');
C = griddata(y,z,c,Y,Z,'cubic');

#### 댓글 수: 3

Philippe Corner 31 Mar 2020
You are amazing man, you really understand deeply the functions. You are helping me a lot with this understanding. Thanks a lot I really appreciate your time.
The only thing here is that we have to "sacrify" plotting with the right coordinates.. and use in the final plot the scaled valies, right?
darova 31 Mar 2020
You can scale them back
sc = (max(z)-min(z))/(max(y)-min(y));
y0 = mean(y);
x0 = mean(x);
y = (y-y0)*sc;
x = (x-x0)*sc;
yv = linspace(min(y), max(y), 50);
zv = linspace(min(z), max(z), 50);
[Y,Z] = meshgrid(yv, zv);
X = griddata(y,z,x,Y,Z,'cubic'); % made a mistake before
C = griddata(y,z,c,Y,Z,'cubic'); % made a mistkae before
Y = Y/sc + y0; % scale back
X = X/sc + x0;
surf(X,Y,Z,'cdata',C) % plot original
Philippe Corner 31 Mar 2020
Oh man, as we say in my country. you are crack! haha you rock man, amazing, thanks!