bin cube: Perl snippet

bin_cube.perl

if ($#ARGV<2) {
    print "USE: bin_cube.pl INPUT.CUBE.fits BIN_BOX OUTPUT.CUBE.fits \n";
    exit;
}

$infile=$ARGV[0];
#scale factor
$N=$ARGV[1];
$outfile=$ARGV[2];
#Perl data language: read FITS file
$pdl=rfits($infile);
$h=$pdl->hdr;
#dimensions of input image
($nx,$ny,$nz)=$pdl->dims;
#dimensions/scale factor; if $N < 1 -- array is enlarged
$NX=int($nx/$N);
$NY=int($ny/$N);
#output array
$out=zeroes($NX,$NY,$nz);
#reading info from FITS header:
#dimensions of input array
$$h{NAXIS1}=$NX;
$$h{NAXIS2}=$NY;
#physical scale
$$h{CRPIX1}=$$h{CRPIX1}/$N;
$$h{CRPIX2}=$$h{CRPIX2}/$N;
$$h{CDELT1}=$$h{CDELT1}*$N;
$$h{CDELT2}=$$h{CDELT2}*$N;
$$h{CD1_1}=$$h{CD1_1}*$N;
$$h{CD2_2}=$$h{CD2_2}*$N;

#loop through the array
for ($i=0;$i<$NX;$i++) {
    for ($j=0;$j<$NY;$j++) {
        #cut out small slice out of the array
	$nx1=$i*$N;
	$nx2=($i+1)*$N;
	$ny1=$j*$N;
	$ny2=($j+1)*$N;
	if ($nx1<0) {
	    $nx1=0;
	}
	if ($ny1<0) {
	    $ny1=0;
	}
	if ($nx2>=$nx) {
	    $nx2=$nx-1;
	}
	if ($ny2>=$ny) {
	    $ny2=$ny-1;
	}
# The section below: it resizes the array and takes the _mean_ of the bin as the rebinned pixel value
#instead of preserving the flux, or am I wrong?
   

	my $sec=$pdl->slice("$nx1:$nx2,$ny1:$ny2,");
	my $sum=sumover($sec); #http://pdl.perl.org/?docs=Ufunc&title=sumover#sumover
	my $sum2=sumover($sum); #summing across the other dimension of the slice
	$t = $out->slice("($i),($j),"); 
	$t .=$sum2/($N*$N); #sum is divided by the number of initial pixels in the bin
    }
}

#write out
$out->sethdr($h);
$out->wfits($outfile);