Trick for Astrometry.net Plate Solving Setup

[JonCarleton]

I struggled with the setup for Astrometry.net from the very first time I got a headache reading the online documentation.  I use Linux (Ubuntu), so I did the "manly" thing and ran:

sudo apt install astrometry.net && sudo apt install astrometry-data-tycho2

All installed without error and I retired to bed thinking, "What a good boy am I."

Next evening, I tried a plate solve with the main scope (via CCDCiel...calling Astrometry via solve-field) and it ran a very long time and failed (timeout).  So I took a picture through the guide scope (same location) and it solved, though it took a very long time and was in danger of timeout.  Next, I sent both images to the online server (https://nova.astrometry.net) and they both solved.  Quicker than the local server for the one that solved locally and successfully for the main scope image that failed locally.

Now, I understand the problem is getting the right index files and matching the right Astrometry image files with your scope parameters for the part of the sky you are looking at and yada-yada-headache-yada.  But I noticed something that led to a shortcut, so I thought I'd share:

When you solve via the website, it =tells= you which index it used....kinda.  In my case, it said "solved using index-205.fits" for both images.  From previous headaches, I know this relates to the 4200 series index files. and so I need all the 4205 files.  So, I did a quick drain flush of the tycho2 data via :  sudo apt purge astrometry-data-tycho2 and went to http://data.astrometry.net/4200/ to download all the index-4205-nn.fits files and placed them where the tycho2 files used to live. In my Ubuntu system:  /usr/share/astrometry.

Results:  Bam!  Near instant plate solve locally on both images.

Now, am I done?  Probably not.  For different parts of the sky for my setup I may need to grab a few more file sets.  Probably the ones adjacent to 4205, such as 4204 or 4206.  It may take me a few plate solve failures to collect all the indexes I need.  But if I get a failure, I have a simple process for resolution:  Solve it online, see which index it used and get that -one- series.

The key to fast solving is not to have ALL the indexes and to have ONLY the ones you need for your scope, location and sky.  I'm pretty sure, for example, I don't need Southern hemisphere stuff for the Southeastern US. 

Why didn't the tycho2 files work (for the main scope image, anyway).  They would have, had I the right ones and ONLY the right ones or if I had set the timeout high enough.  Too many index files will slow the process significantly.  I just didn't have such a simple method of identifying the index I actually needed with the tycho2.  Although, through the fog of previous headaches, I do recall seeing a rather complex method for determining the ones you need from other index series.

[callump]

There is a table in the Astrometry.net README that gives the range of the skymarks for each set of index files. If you know the size of your field, you can just use the height in arc-minutes to work out a range of index files you need. E.g. for a 30 arc-minute image you might want to use:

index-4205-*.fits    11–16
index-4204-*.fits    8–11
index-4203-*.fits    5.6–8.0
index-4202-*.fits    4.0–5.6

/callump