Automatically fit grid spacing

Hello everyone,

I'm looking for a smart/fast way to automatically generate a fitting grid spacing. I have a set of x-y-z points that I plot with the meshgrid function as shown in the following code where x and y are vectors with the x/y values of the points. fitresult is a sfit generated by the curve fitting tool box.

tx=min(x):spacingX:max(x);
ty=min(y):spaxingY:max(y);
[xm,ym]=meshgrid(tx,ty);
zm=fitresult(xm,ym);
mesh(xm,ym,zm)

spacingX and spacingY are to be the automatically generated values for my grid spacing. So basically what I need is a code that takes a look at max(x)-min(x) and then decides what grid spacing would give an accurate fit.

For example I manually use a spacing of 0.001 for a range of min(x)=0.0194 : max(x)=0.2353. The y-values for the same data set range from min(y)=8.5928 : max(y)=119.5933, so using the same spacing is impossible.

Is there a matlab function or mathematical way to automatically give me a fitting grid spacing?

Thanks for your help!

Answer 1

[KSSV]

Note that either you have to provide grid spacing or number of points you want along a axes.....if you fix....grid spacing..number of points are decided. Else if you provide number of points grid spacing is decided.

%%Number of points given 
x = linspace(min(x),max(x),N) ;
%%Grid spacing given 
x = min(x):dx:max(x) ;

[Ingo Rück]

So I could basically say that N=2*length(x) to have the spacing somewhat depending on my data. That doesnt seem very consistent because some data sets have outliers that would stretch the grid and make the mesh less accurate..