Creating a function (pathfinding algorithm)

How do you create a function:

function [ filled ] = travelDistance ( blank )

where: blank is a two-dimensional array comprised of -1s, 0s and 1s, and filled is blank modified to the specifications below.

To create filled, replace every 0 in blank with its distance to the nearest 1 (staring at 2), traveling along cardinal directions (up, down, left and right) without passing through a -1 value. That is, all 0s that are directly next to a 1 should be changed to a 2 and all 0s next to any new 2s should be changed to 3s, and so on. Spaces with -1 values should be treated as “walls” and should stay -1. If there is no route from a 0 to a 1 without passing through a -1, it should remain a 0. The ultimate result should be that every 0 that is connected to a 1 is replaced with its distance to the nearest 1.

[KSSV]

Okay..what you have attempted?

[amateurintraining]

I tried doing this:

function [ filled ] = travelDistance( blank )
%TRAVELDISTANCE
%   blank: two-dimensional array comprised of -1s, 0s, and 1s
%   filled: blank that is modified (replace every 0 in blank with its
%   distance to the nearest 1, tarting at 2, traveling along cardinal
%   directions without passing through a -1 value)

    function d = xyspace(A,x,y)
        idx1=find(ismember(A,x));
        idx2=find(ismember(A,y));
        x=union(idx1,idx2);
        whichset=ismember(x,idx1)+2*ismember(x,idx2);
        idx=diff(whichset)>0;
        d=x([false,idx])-x([idx,false]);
    end

filled=xyspace(blank,0,1);

end

But this results in an error:

Error using horzcat
Dimensions of matrices being concatenated
are not consistent.
Error in travelDistance/xyspace (line 14)
      d=x([false,idx])-x([idx,false]);
Error in travelDistance (line 17)
filled=xyspace(blank,0,1);

I think this error occurs because the function only works for a 1xn array.

Answer 1

[Image Analyst]

To do it that way (manually), then fist compute the distance transform with bwdist().

Otherwise, do it automatically with bwdistgeodesic(). See Steve's blog: http://blogs.mathworks.com/steve/2011/11/01/exploring-shortest-paths-part-1/