views:

13

answers:

0

I know kd-trees are traditionally used to store points, but I want to store lines instead. Would it be best to split the line at every intersection with the splitting of the kd-tree? or would storing just the end-points into kd-suffice for nearest neighbor finding?