Astrometry Experiment

[steppenwolf]

I am currently involved as an alpha tester on a project to build an 'astrometry engine' to create correct, standards-compliant astrometric meta-data for every useful astronomical image ever taken, past and future, in any state of archival disarray.

The astrometry engine will take any image and return the astrometry world coordinate system (WCS) - i.e., a standards-based description of the (usually nonlinear) transformation between image coordinates and sky coordinates, with absolutely no 'false positives' (but maybe some 'no answers'). It will do its best, even when the input image has no meta-data or totally incorrect meta-data.

The project engineers intend to install the engine for real-time operation on the web, at observatories, at plate-scanning projects and at data archives.

The following is an example of the output generated by the engine from an image I took in August this year with my SXVF-H9C with 50mm f2.8 lens stopped down to F5, piggy back mounted on SkyScan EQ6. 8 images of 200 secs. each stacked in MaximDL. Guided.

The Image Field is at RA,Dec = (315.13, 45.0823) degrees

and spans 9.94 x 7.43 degrees .

The field contains:

* IC 5068

* IC 5070 / Pelican nebula

* IC 5076

* NGC 6997

* NGC 7000 / North America nebula

* NGC 7039

* The star Deneb (?Cyg)

* The star 55Cyg

* The star 56Cyg

* The star 57Cyg

* The star ?Cyg

* The star 59Cyg

* The star 60Cyg

* The star ?Cyg

* The star 63Cyg

* The star 68Cyg

[GazOC]

Impressive Steve, I'm sure (I know!!) I've seen this somewhere else over the last week or so......

[themos]

This is the opposite of GoTo, isn't it?

[steppenwolf]

This is the opposite of GoTo, isn't it?

What, a sort of 'Gone To' as in 'where have I gone to?' 8)

[cygnusx1]

I could defo use that, esp when imaging small faint objects and trying to get the object properly placed in the fov of the camera.

Having some star/ngc indicators would help immensely.

Cost, availability?

[steppenwolf]

Cost, availability?

The engine will be made freely mirror-able anywhere by anyone for download and the originators will provide their algorithms and code to the public. Additionally, they intend to install the engine for real-time operation on the web, at observatories, at plate-scanning projects and at data archives.

[beamer3.6m]

Sorry, I may be thick, but what does all this mean?

[steppenwolf]

Sorry, I may be thick, but what does all this mean?

What it means is that in the future, any image of the night sky taken by an amateur or a professional (even images taken years ago), can be analysed by the astrometry engine and the exact location of the captured data in the night sky determined very accurately. This can be very handy at a simple level by indicating what it is you have captured but a whole load of other data will also be available to the imager and all based on a simple astro photograph. The sample above had NO positional data supplied with it and was just a standard .jpg but having submitted the image to the engine within a minute or so I received a wide range of data about it's exact location, orientation, size and what major objects had been captured in the frame.

To me this is pretty cool stuff but then I listen to 'Nightwish' so my judgement may well be flawed :?

[saturn5]

Nothing wrong with your judgement or Nightwish,Dark Passion Play is my fav CD of the year.

That makes me a 48 y.o goth

Btw ,that will be one mega impressive search engine.

[steppenwolf]

Dark Passion Play is my fav CD of the year.

How funny, mine too but do you prefer the 'instrumental' of the 'voice' version best? Track 11 is the same in any case!

[saturn5]

I prefer Anette singing,Tarja whilst having an impressive operatic voice was beginning to sound a bit samey and predictable so i like the new line up better,as for the instrumental or vocal version,if i just want to sit back and relax its the instrumental version(as loud as the headphones or my ears can take)symphonic rock i think it is now called,i think movie score metal sounds better.

Poet and the Pendulum is my fav instrumental one and Sahara is my fav vocal.Ican now play track 11 on the guitar thanks to my 15 y.o son showing me how.It took him all of 10 mins to show me what had taken me about 2 hrs to work out,(didnt tell him that though).

One final thing,i have listened to the instrumental version whilst stretched out on the sun lounger looking up at the stars,the slower moody tracks are just made for listening to under the stars,must be that nightwolf thing or something. :evil2:

Pink Floyd's Shine on you crazy diamonds is another good stargazing song.

[steppenwolf]

symphonic rock

Hmmm, I prefer 'Viking Battle Metal' - helps break the ice at parties - just before I drop the bombshell that I do Astronomy ...

[saturn5]

How true the never judge a book by its cover saying is.

One minute your cool dude listening to goth,symphonic,viking metal and then its,"oh,your into astronomy,thats err mmm,really eh cool,what is your star sign?"

[saturn5]

I'd just like to add that it was your website that swayed me to get the Skywatcher 250,i had been unable to make up my mind what i wanted until i saw the picture of you standing behind your scope,i said to myself,i gotta have one of those.You also did an excellent review on it,cant remember where but it came up in a google search when i was desparately needing guidance.

I have been trying to get my head around the astrometry engine and i find the whole concept quite mind blowing to say the least(excessive listening to loud goth metal damages brain)

[FLO]

... any image of the night sky taken by an amateur or a professional (even images taken years ago), can be analysed by the astrometry engine and the exact location of the captured data in the night sky determined very accurately.

Blimey :shock:

Now that would be impressive 8)

[steppenwolf]

Now that would be impressive

Steve, I tried several images to 'fool' the system (that is what alpha testing is for, of course!) and this one of Comet Holmes could have put a spanner in the works but as you can see it's location has been accurately confirmed - I am impressed already!

[steppenwolf]

I'd just like to add that it was your website that swayed me to get the Skywatcher 250,i had been unable to make up my mind what i wanted until i saw the picture of you standing behind your scope,i said to myself,i gotta have one of those.You also did an excellent review on it

I hope you are as pleased with you 250 as I am with mine!

[winfij]

I am currently involved as an alpha tester on a project to build an 'astrometry engine' to create correct, standards-compliant astrometric meta-data for every useful astronomical image ever taken, past and future, in any state of archival disarray.

That's a bold claim!

I've seen this new tool mentioned a number of times recently, but noone seems to be able to answer whether it's going to work with the small fields of view normally involved with amateur scopes (< 30 arcmin)

It's relatively easy to identify the target with a widefield shot where you can look for patterns in the relatively uncommon very bright stars, but when the brightest star in the image is down less than mag 10, there's a heck of a lot of them in the sky to compare.

I've used PinPoint extensively and it struggles hard on images which are even a couple of image widths off from their stated centers.

If this really does work for fainter fields, it'll be a great replacement for PinPoint for automated astrometry if someone ports it to a dll.

John

[steppenwolf]

That's a bold claim!

I've seen this new tool mentioned a number of times recently, but noone seems to be able to answer whether it's going to work with the small fields of view normally involved with amateur scopes (< 30 arcmin)

Luckily, it's not my claim - this is a quote from their website so I have no axe to grind! However, I have submitted quite a few JPEG images to them and all but one have been solved immediately (the one that didn't solve was an appalling image sent on purpose to press the system - a scan of an old film 6 x 4 film image!). The following are the dimensions of two of the solved images I sent them and the actual images themselves:-

25.80 x 19.28 arc minutes

25.67 x 19.18 arc minutes

As you can see, they comfortably fit your criteria for an amateur 'scope image. Bold claim it may be but it is an impressive system and I am not including any positional detail with my JPEGS so the system is working for it's keep on my uninspiring images!

[winfij]

That's excellent then - those certainly do fit my criterion.

I look forward to taking a look at the source/algorithm when they release it!

John

[cygnusx1]

SW, i've had a look round the site, can't find a link where to submit images for analysis, only Unix source d/l?

[steppenwolf]

can't find a link where to submit images for analysis, only Unix source d/l?

You have to be accepted as an alpha tester first. Have a look at the second paragraph headed 'WEB' under the 'USE' tab for details on alpha testing.