I put together a piece of code that might help you:

#!perl -sw

my $input = [[1, 3, 4, 2],
             [7, 5, 3, 6],
             [5, 4, 2, 4]];
my $SEP = " ";
my $send;

sub printarray {
   my $arr = shift;
   for my $t (@$arr) {
      for (@$t) {
         print "$_ ";
      }
      print "\n";
   }
}

sub encode {
   my $arr = shift;
   join $SEP x 2, map { join $SEP, @$_ } @$arr;
}

sub decode {
   my $in = shift;
   [(map { [(split $SEP, $_)] } split($SEP x 2, $in))];
}

printarray($input);
print "\n" x 2;
$send = encode($input);
print $send . "\n" x 2;
$input = decode($send);
printarray($input);
[download]

It seperates the elements of one line by one separator and two consecutive lines by two separators. Note that although I embedded everything in subs (a nasty little habit of mine) the essential parts are one-liners.
encode expects a reference to an array, decode returns an unnamed array.

Hope this helped.