I'm not sure how to get that information in a pure perl way. But the following works great on my Linux system:

$_ = (`df -k /home`)[1];
my $blocks = (split)[1];
print "$blocks\n";
[download]

The first line of df output is column headers, so it skips that and takes the second line. Then it splits that on whitespace and takes the second token, the block count.