NavList:

   

Reply

I was playin with apps on my phone that give a compass graticule
overlaid on an image from a front facing camera, sometimes, Lat and
Long, a map etc as well. Neat, I thought, this is the new "monocular
compass" (well you can also stick on a monocular but my camera has a
telephoto lens anyway).   But some of the apps I tried only worked
with the phone horizontal. I presume they misread the output from the
magnetometer.

That had me thinking. Suppose you point a camera upwards and have a
good compass to orient it. That would give you the azimuth of
astronomical bodies near your Geographical position. So some quick
calculations (open to refinement). Suppose you observe with six
degrees of arc of your Zenith and get the azimuth to an accuracy of 1
degree (maybe OK with enough averaging and stabilization).

The density of "visible" stars (for some value of visible, lets say a
clear night away from light pollution) is around 0.1 to 0.2 stars per
square degree I think. Well it actually varies a lot but suppose this
hypothetical device is sitting there watching the sky all the time
with nothing better to do and sometimes gets lucky with lots of stars
near the Zenith.  So 6 degree FOV radius is about 360 square degrees,
so we will have between 3 and 7 starts to work with typically. So we
have position lines as though we took bearings on the GP of  the stars
and work as though it was a hand bearing compass fix  on objects up to
360 Nautical miles away. A one degree error is about 6 nautical mile
as 360 nautical miles, but that is the extreme and much of the time
there will be stars closer to the GP.  If we average over time
(and have good accelerometers and DR) and combine more than 3 LOP, etc
we can perhaps do as well  as a round of sextant sights, but with a
less accurate horizontal needed easier to automate.


So a few comments.

This is not even back of the envelope. I just thought of it driving so
I may well have made some mistakes.  I usually do. Even if I did for
some star density, accuracy etc it  would work in theory. Feedback
welcome.

While it would be handy to have a giro stabilized camera looking up,
we could also pattern match the visible stars and average as the field
of view moves around, maybe a wider field of view. Presumably the
wealth of astrophotography algorithms would help.  If I am right it
needs an approximate vertical but not a horizontal to minutes of arc
as is traditional.

Typically on ocean crossings one fix per 24 hours is fine. This could
give quite a few fixes in the night, and could be automated.

Presumably either this has been tried and didn't work but maybe not
with the current technology. One can buy an automated deep sky
astrophotography set up for a few hundred pounds now (eg ZWO Seestar),
and computation can often be a substitute for more expensive optics.
You can also get automated camera gimbals cheaply, beloved of the
social media vloggers.

Looking forward to hearing your thoughts and corrections.

Bill Lionheart

   

Reply