2d extrapolation to replace NaN in corners

Hi !

I'm trying to do 2d extrapolation for my table in matlab.

I have a matrix with 101x101 values and corresponding axes with 1x101 and 101x1 values.

Now my problem is the following:

The corners of the matrix are containing NaN values which I want to replace with extrapolated values of the given data.

Answer 1

[Matt J]

One option is fillmissing.

There are also offerings on the File Exchange, e.g.,

https://www.mathworks.com/matlabcentral/fileexchange/4551-inpaint_nans?s_tid=ta_fx_results

[Leon Thürnau]

Hi ! Thanks a lot for your reply !

I allready self programmed something which is just performing like, so at least I learned that I could have saved time...

Mfm = fillmissing(M, 'nearest')

Unfortunately this isn't exactly what I want, I'm looking for a kind of interpolation / extrapolation of the given data.

Argument "linear" and "spline" aren't working as expected, but linear is the closest to what I need.

Because of saturation the curve should become flatter without rising again.

[Leon Thürnau]

Update: "Linear" with dim = 2 seems so solve my problem !

[Matt J]

Glad to hear it, but since fillmissing worked, please Accept-click the answer.

Answer 2

[Star Strider]

The fillmissing function may do what you want.

Try something like this —

M = NaN(5);

M([3 (7:9) 11:15 (7:9)+10 23]) = randi(9, 1, 13)

M = 5×5

NaN NaN 4 NaN NaN NaN 6 6 2 NaN 2 6 1 6 4 NaN 5 9 4 NaN NaN NaN 2 NaN NaN

Mfm = fillmissing(M, 'nearest')

Mfm = 5×5

2 6 4 2 4 2 6 6 2 4 2 6 1 6 4 2 5 9 4 4 2 5 2 4 4

.