Have you seen URI::Find?

Otherwise, please post a short, self-contained program and representative input data. I suggest you review perlre about greedy repetitions. .* will always take up as much as it can while the regular expression still matches. Also, think about whether you really want to use the dot - it is unlikely that whitespace delimits an URI.

Also, your usage of $1 and $2 is weird because you never capture any data.