Plotting locations...

[discardedastro]

I'm lucky enough to have a reasonably large garden and field near the house, so I'm slightly spoilt for choice on where to put an observatory, though there's plenty of constraints for each potential site.

I wanted to get a good set of horizon profiles - I've seen some of the makeshift horizon visibility measurement tools, but this seemed pretty slow to do and coarse, not to mention fiddly to digitise and compare. I had a bit of a brainwave and figured I'd share the results.

I have a 360 degree camera, a first-generation Ricoh Theta S - you can get them on eBay for sensible money. Alternatively, you can take regular photos (carefully, with a levelled tripod) and stitch them in software for this purpose. The Theta spits out equirectangular-projected images which are stitched in the camera for an approximately 360 degree view, using two lenses/sensors. So I figured all I needed was to take nice level photos, north-aligned, and I could do some simple software to draw a horizon line. This is the sort of photo you get out (once I've blurred out my ugly mug):

Of course, I forgot to north-align my photos when taking them. While it would, as my partner pointed out, have been quicker to go re-take the photos I figured I'd have some fun so I wrote a rudimentary moon-spotting algorithm (using prior knowledge of approx location+time of the photos), used the moon azimuth to calculate which pixel column was north, and offset everything by that. Deeply unscientific, but for a relative comparison, good enough to get everything pointed roughly the same way!

My horizon detection is very simple, and just takes a ~4 degree block of pixels, looks for darker pixels, and calculates an average height, the blue dots in this:

The script outputs a CSV file with a set of azimuth/altitude values which can then get dropped into pandas/matplotlib or Excel for some simple plotting:

The script is written in Python and can be found here, for those who feel inclined to tinker: https://gist.github.com/JamesHarrison/fd75fb768d0825d3a9b4db5622656f1b - it requires the skyfield, numpy, pandas and OpenCV Python libraries, and is fairly well commented.

Not quite sure it's given me a good answer on where to put my future obsy, but it's some more data! I think I'm going to look at some further analysis where I take some common target catalogues and calculate aggregate visibility over a year for all the sites...

Slightly mad, but it's a better way to pass the time than staring at my bank account balance and hoping it moves.

[Howie_Oz]

Decisions, decisions! LOL.

And .... if you raised the OBS one or even two metres off the ground all the above would change again! LOL

All the best mate. Can't wait to see where you decide. Post up photos as you build ... cheers

 

[Gina]

Ooooh another obsy build on the way - yippee    Bring it on...

[discardedastro]

11 hours ago, Howie_Oz said:

Decisions, decisions! LOL.

And .... if you raised the OBS one or even two metres off the ground all the above would change again! LOL

All the best mate. Can't wait to see where you decide. Post up photos as you build ... cheers

 

Yeah - I took all these at roughly the saddle height of my parked EQ6-R Pro, thinking I don't want to go much higher than that (though will add in another 30-50cm or so undoubtedly). The other thing I wanted to look at was, given this, how large a footprint I need versus how high the walls will be. Side walls I can have partially on the roll-off, of course, but the end wall/warm room not so much. But this does suggest that in pretty much all my locations I have a ~45-60 degree arc where I'm going to get no better than 30 degrees visibility anyway, so I can use that as a design target for the ends and minimise my loss of visibility due to structure.

Will set up a thread once I get to actually building!