PixInsight Forum (historical)

PixInsight => Tutorials and Processing Examples => Tips & Tricks => Topic started by: dsnay on 2011 January 05 04:44:41

Title: Image Registration of DSLR data
Post by: dsnay on 2011 January 05 04:44:41

Hi,

Is there some change I could make to the settings in the Image Registration window which would speed up the process of registering the monstrous files produced by DSLR's without negative impact on the results? I don't understand the various fields in there well enough to make an educated guess.

Thanks,
Dave

Title: Re: Image Registration of DSLR data
Post by: sjbrown on 2011 January 05 06:39:45

Hi Dave,

You could reduce the Maximum stars setting under the Star Matching bar from the default (Auto) of 2000.

When I run a subset from a given night I reduce it as low as 200 stars with excellent results and a huge
cut in the processing time. For production runs of multiple subsets I leave the default setting just to ensure
quality of the alignment.

Hope this helps,

John

Title: Re: Image Registration of DSLR data
Post by: dsnay on 2011 January 07 09:22:49

Thanks John.

Title: Re: Image Registration of DSLR data
Post by: dsnay on 2011 January 11 04:46:17

Quote from: sjbrown on 2011 January 05 06:39:45

Hi Dave,

You could reduce the Maximum stars setting under the Star Matching bar from the default (Auto) of 2000.

When I run a subset from a given night I reduce it as low as 200 stars with excellent results and a huge
cut in the processing time. For production runs of multiple subsets I leave the default setting just to ensure
quality of the alignment.

Hope this helps,

John

John,

When I expand that bar, the only field with a value of 2000 is the Ransac Iterations field. The Maximum Stars field is set to "Auto". Should I be setting that field to 200?

thanks,
Dave

Title: Re: Image Registration of DSLR data
Post by: sjbrown on 2011 January 11 07:53:26

Yes, the "auto" setting usually equates to  a value of 2000 stars for my images (12 MP on a Canon 450D). The value of
200 is an arbitrary value but it still produces very good results.

John