You could also just use a match global to extract all 5 digit numbers that are preceded by a space.

#!/usr/bin/perl -w
use strict;

my $text = 'scaffold_12147 protein_coding exon 32192 32242 . - . gene_
+id "ENSTBEG00000007763';

my @numbers = ($text =~ /\s(\d{5})/g);
print "@numbers\n";

#prints 32192 32242
[download]

Update: slight wrinkle:

#!/usr/bin/perl -w
use strict;

open FILE, "vik.txt" or die $!; 
while(<FILE>)
{
  my ($exon, $number1, $number2) = (split /\s+/, $_)[2,3,4];
  if (defined ($exon) and $exon eq "exon")
  {
     print "$number1 $number2\n";
  }
}
[download]