distance of points to points calculation

Hi: I have two set of 2D scatter points, says A and B, I need the minimun value of each points in A to B. for example, the size of A and B are 100*2, and below are my codes:

      for i=1:1:100
        ppdist=((A(j,1)-B(:,1)).^2+(A(j,2)-B(:,2)).^2).^0.5;
        result(i)=min(ppdist);
      end

however, if the size of A and B are significantly large, for example: 1 million for each, the calculation cost significantly large amount of time.

so I want know if there is anyway to optimize this calculation?

Thanks Yu

Answer 1

[Anton Semechko]

Use 'knnsearch' function. For example:

 [idxA,D_a2b]=knnsearch(B,A,'k',1);

will return indices of the points in set A closest to B, so that A(idxA(j),:) is the closest point to B(j,:). D_a2b(j) is the distance between these closest point pairs.