> I don't think GAIA is doing anything other than calling autoastrom which retrieves the catalogue.
>
Indeed, this is correct.
>
> The documented interface for --obsdata (see "perldoc Starlink::Autoastrom") suggests decimal hours for RA and decimal degrees for dec.
>
> The problem is that Starlink::Autoastrom has a bug in it around line 874:
>
> if( $key eq 'DEC' ) {
> # Convert to Astro::Coords::Angle object.
> if( $value =~ /^[+\-]?\d+:\d+:[\d\.]+$/ ) {
> $value = new Astro::Coords::Angle( $value, units => 'sex' );
> } elsif( $value =~ /^-?[\d\.]+$/ ) {
> $value = new Astro::Coords::Angle( $value, units => 'hour' );
>
> so that's a cut and paste error from the "RA" section. Should be "units => 'deg'". Astro::Coords::Angle does not know what to do with hour so ignores it and uses the default of radians.
>
Fixed in the repository
> is this because there are too few stars?
>
Could be an option, but I just manually inserted an astrometric
calibration on the test image, using 22h 41m 28s.5, -3d 43m 48s and a
scale of 0.625 on both axis, and overplotted a USNO catalog: there is no
obvious match between the two.
Luca
|