Finding 4 closest points

Hi,

So I am trying to use bilinear interpolation to extrapolate data for certain values of a 3-dimensional array. But, the issue is that the bilinear interpolation requires an input of the 4 closest numbers by distance in order for the equations to be solved. Does anyone have an idea of how I can obtain these 4 values. I am going to put some context below for better understanding.

62 0 63 0

0 0 0 0

63 0 61 0

0 0 0 0

I am going to refer to the coortinates of the elements in the form of (row,column).

So for example, consider this array above. For context, I need a way to find the 4 closest non-zero points to (1,2). If I solved this manually, the points would be (1,1),(1,3),(3,1),(3,3).

This is only a small part of a very large array that I have given for context. If anyone can help me with the coding, it would be appreciated.

[Jonas]

what aboit using find() with two output arguments which gives you the x and y index of the nonzero elements. after then you can calculate on the base of this indices the distance to the locations you are interested in.

Answer 1

[Rik]

You can design something yourself, but you could also see if fillmissing solves your problem. The documentation suggests it should work for 3D arrays as well.