Re: Satellite Elements from Astrometric Observation.

From: John Burns (seesat@john.geek.nz)
Date: Wed Mar 18 2009 - 12:48:25 UTC

  • Next message: Bruce MacDonald: "RE: Reading list for new observers"

    I've been playing with this software over the last few hours.
    It seems reliable and somewhat accurate.
    
    It takes anywhere from 15 to 60 seconds to solve an image and relies  
    on up to 15GB of indexes to match against - most images of 10degree or  
    wider field of view only require approx 30 MB of indexes.
    
    I've been in contact with one of the developers and it appears that  
    given an image, we should be able to get back pixel X and Y  
    coordinates with a matching J2000 Ra/Dec for every matched star in the  
    image.  This would make detecting the Ra/Dec of the satellite trivial  
    - perhaps by zooming in and clicking on the relevant end of the object.
    
    The best integration for this would be web based with a client  
    application firing off the images to the server for processing. The  
    client app could batch upload images and then get the results once  
    processed.
    
    For any given image I would expect the time required by a human to  
    accurately log the observation to be the order of less than 1 minute  
    per image (not including queue times for processing the images on a  
    remote server.
    
    Unfortunately for me, the software runs under linux and I don't have  
    any spare linux servers available right now.
    
    If anyone has knowledge of Linux and preferably has a server with a  
    little spare capacity (and wants to help) I'd be grateful for  
    assistance to get something working.
    
    John
    
    
    On 17 Mar 2009, at 15:23, George Roberts wrote:
    
    > Just recently (a few weeks ago) the folks at google have released  
    > free software that will analyze any photograph taken of the night  
    > sky and tell you it's position/orientation.  The web interface is  
    > still in alpha and isn't yet available.  Source code and indexes are  
    > also available but the indexes are only available for non-commercial  
    > use.
    >
    > If John, Andrew or someone else wants to use this software for  
    > seesat-l purposes -- and if their application is for anyone to use,  
    > then more details are here:
    >
    > http://astrometry.net/use.html
    >
    > - George Roberts
    >
    >
    
    
    -------------------------------------------------------------------------
    Subscribe/Unsubscribe info, Frequently Asked Questions, SeeSat-L archive:  
    http://www.satobs.org/seesat/seesatindex.html
    



    This archive was generated by hypermail 2b29 : Wed Mar 18 2009 - 12:49:38 UTC