You should maybe look into URI::Heuristic for completing your links. Something like this:

#!/usr/bin/perl
use strict;
use URI::Heuristic qw(uf_uristr);
my $url = uf_uristr("www.perlmonks.com");
print "<a href=$url>$url</a>\n";
[download]

If the user input is www.perlmonks.com or http://www.perlmonks.com it will give you the proper URI. In this case with html code:
<a href=http://www.perlmonks.com>http://www.perlmonks.com</a>
using www.perlmonks.com as input.
It also does ftp and others.

ok..it's a bit off what you really asked for but i just noted this module at CPAN and liked it ;o)

Update:

#!/usr/bin/perl
use URI::Heuristic qw(uf_uristr);
use strict;
my $url;
my $input = "<a href = http://www.perlmonks.com>perlmonks</a>";
#my $input = "www.perlmonks.com";
if ($input =~ /^</) {
    $input =~ /<a\s+href\s*=\s*(.*)>.*<\/a>/i;
    $url = uf_uristr("$1");
} else {
    $url = uf_uristr("$input");
}

print "<a href=$url>$url</a>";
[download]

This works just fine, also with resolving the url to an uri. You may also drag out the text by enclosing the last .* and drag out the value of $2.
The regex should work on it's own without using the module also, me thinks.