Upsampling the data using interpolation gives NaN values.

Question

Ajay Kumar 2019-9-26

编辑： Ajay Kumar 2019-10-7

Hello everyone,

I have the data of size 2173x1. I want to increase the number of samples to length of 2225570x1 using interpolation/resampling.

I tried using interp1 function and ended up having the required length, but NaN values after 2173 samples.

Thanks in advance.

Answer 1

Star Strider 2019-9-26

Use the 'extrap' argument in interp1.

Example —

D = load('x.mat');
x = D.x;
v = 1:numel(x);
ve = 1:2225570;
x_extended = interp1(v, x, ve, 'linear','extrap');

This is dangerous, because you have no idea what ‘x’ actually would be beyond its current definition. I would definitely avoid that.

However you can completely avoid all those problems if you simply want more points in ‘x’:

vr = linspace(min(v), max(v), 2225570);
x_resampled = interp1(v, x, vr);

Experiment to get the result you want.

Ajay Kumar 2019-9-26

Yes, I have tried 'extrap' before, and that goes beyond the maximum limit of original data and I simply neglected it.

But anyways, the linspace thing worked out. Thanks :)

Star Strider 2019-9-26

As always, my pleasure!