Anyone got tips on what software is available to aid supernova discovery with respect to processing/analysing of captured data? I have CCDSoft which has functions to do this, but it relies upon plate-solving which is often a problem for my setup. I'm looking for something which can align and then compare two sets of images, reporting appropraite differences.
Maxim DL contains probably the best blink comparison tool I have seen. I believe CCD Soft can also blink images. The Maxim Supernova blink programme aligns a reference (say your own or a DSS image) image with your current image, you can set the blink rate but you still need your eyes to note any changes. It really is very simple to use and quite effective (Stu Parker has found 11 this way) but is laborious if you have hundreds of images.
We have looked into image subtraction software but I think this is difficult to find at the amateur level.
If you want a much cheaper blink tool then you can't go past Astrometrica. (and I have MaxIm and I wouldn't call it the best blink comparison tool). At least with Astrometrica it incorporates a minor Planet overlay so you can immediately rule that out as the 'source'.
When you are checking possibly a hundred plus images, both speed and accuracy are required so in my opinion Maxim wins hands down.
In Astrometrica as I understand it, there is no auto rotation of the image or image scaling - this is time consuming, Maxim does this automatically.
Plate solving and plotting asteroids can waste valuable time and is not needed when checking hundreds of images (unless you have all day to do the blinking, rotation, image scaling and plate solving) this is something quite easily done at the time a suspect object is found, there is much to rule out before reporting anything.
The latest mpcorb or astorb data plot some 540,000 minor planets, you can't rule out an unknown minor planet (or variable star known or unknown). There is no easy way around this, we have looked and it comes down to using the best tools available which at this stage still very much includes using your eyes.
It surprises me that there is no image subtraction software available to amateurs for this sort of work, if anyone knows of any please let me know.
My own discoveries thus far have been a direct comparison with my reference images, no blinking just real time eyeballing. If you decide to blink the images it may open up other problems - if you leave them until next day to check then you may lose your discovery to others ie CHASE etc. You also only have one image and under the new reporting guidelines this is not acceptable (can't even get it on the unconfirmed page) so you will then have to wait to get another image, more time lost, yet something you could have done straight away with the successful (for me anyway) but lesser numbers real time eyeball method I use.
PeterM.
So how do you know your SN is not an asteroid? If you are imaging with the same equipment what do you need plate rotation and scaling for? The overlay does not get called unless instigated by the user and you can choose not to solve for moving target detection!
Unless your SN is mag 22 or fainter then you have a very strong case for not being an Asteroid and most amateurs doing SN searches will be luckly to detect a mag 19 or fainter SN....
Maxims auto rotate and scale does not always work either. I have plenty of examples where it fails dismally. Platesolving is just as quick as MaxIms alignment routines (at least on my systems we are talking less than 1-2 seconds) and platesolving is perhaps just as accurate and reliable as Maxims auto align routines. Perhaps the difference is scripting....
But how many amateurs out there have automated systems bouncing from one target to the next? If they use an automated system then the image should be platesolved already to verify that they were pointed at the right part of the sky!
I don't see how image subtraction is any better than blinking in any case. In amateur equipment and amateur skys you are going to get more variations from sky quality than software can do for image subtraction resulting in false positives. Blinking will make it glaringly obvious that sky quality needs to be considered.
So where does science end and a race begin? Is it about doing the science and making or assisting in a discovery or just getting the credit? You make 1 discover or 100 - you still have made a discovery.....
Geez David you have really hoed into me here (actually some great points raised) I responded to a valid question with my thoughts and experience, which by the way have been quite successful.
I don't want to get into a tit for tat response to all your points here, but I must note a few things.
[QUOTE=higginsdj;668497]So how do you know your SN is not an asteroid?
Well its not even a SN its just a suspect/possible SN even after it has been discovered and reported, until it is professionally followed up.
Simply keeping the current astorb data in say Sky6/X can eliminate most asteroids and then following procedures as set out by CBAT (mandatory check of minor planet checker) should rule out any known asteroids, but you already knew that l'm sure. Out of 15 discoveries BOSS have reported on not one known asteroid false alarm, we do our homework really well.
[QUOTE=higginsdj;668497]If you are imaging with the same equipment what do you need plate rotation and scaling for?
Its not always the same equipment, BOSS team use several setups, alt az included as well as using each others reference images and DSS images as required
[QUOTE=higginsdj;668497] Maxims auto rotate and scale does not always work either. I have plenty of examples where it fails dismally.
Yup software is not always perfect, horses for courses, Stu is the best example of it working very well with his 11 discoveries. On sheer numbers though it performs very well.
[QUOTE=higginsdj;668497] I don't see how image subtraction is any better than blinking in any case. In amateur equipment and amateur skys you are going to get more variations from sky quality than software can do for image subtraction resulting in false positives. Blinking will make it glaringly obvious that sky quality needs to be considered.
That's a very good point, perhaps why it's not available.
[QUOTE=higginsdj;668497] So where does science end and a race begin? Is it about doing the science and making or assisting in a discovery or just getting the credit?
Interesting comment. We are amateurs first and foremost, no more or better than you or anyone else reading this. We don't do the science, how could we? Its a collaboration that in our case has led to a professional at one of the worlds largest observatories willing to do follow up work because it forms part of their professional work. I have found professional astronomers are generally very appreciative of the dollar outlay and time outlay (many hundreds of hours per year) amateurs do put in.
A race, perhaps in some ways, it sure is competitive. Large robotic scopes (LSST etc) are soon to dominate and amateur discoveries will be made using a mouse and someone else's image on the computer screen. How sweet it is to say I discovered that with my own telescope/camera.
We will assist anyone who needs it with follow up for a possible supernova discovery. David you are welcome to ring me at anytime of the night, all we ask is that you have done the basic groundwork first.
Oh and my bloody oath the credit is very, very important, if you discover it you deserve it, a day can make all the difference. BOSS ensure it is shared equally with those who contributed to it and this is something very few do.
I think we have strayed from the original question, just use the eyeball method on lots and lots of images and you will be successful.
Ooops nearly forgot a PS - David have a really great Christmas to you and your family and may 2011 be filled with clear nights, all the best!
PeterM.
Interesting read, don't let it get too serious guy's
Only about 50% of my images successfully plate solve, hence my original enquiry for non-platesolve dependent methods. They don't plate solve due to small FOV (and to a lesser degree star brightness). My images are all captured in an automated fassion and most objects are central enough in the image to still use even if the platesolve is a failure.
I hadn't thought of Astrometrica. I have a license but always forget about the package
The subtraction method is of interest to me as purely a way to automate the detection rather than eyeballing the images. I would have thought it would be possible to use various normalising, de-noising, etc methods to get a meaningful subtraction result. But, not volunteering to write the software myself I can understand there's likely a good reason it doesn't exist for us already.
I'm not much concerned about minor planets being in the FOV at this stage, but can see the potential for having them "get in the way". I think step one is for me to find what I think is a nova before being concerned if it really is one
Blinking is looking the most viable option at thi stage. I already have reduction and alignment of images automated so have realised that even though it would be a bit time intensive I can open up the sets of aligned images and blink them without going through CCDSoft's supernova/minor planet routines which require platesolving.
Maxim has a full download 30 day trial (about 27 meg, see link below) . The cheapest version that contains the SN search tool is $299 US for download, you might pick one up on Astromart or Cloudy Nights much cheaper, they do pop up. Worth running the free trial and seeing how it may be of use. Our experience with blinking sometimes 500 plus images has shown this is the best of what's available.
I wish you and your family a great Christmas.
I hope you have great success with SN hunting and if we can be of any assistance (we have heaps of reference material) please let me know by pm. Actually with you on the otherside of the country to most of the BOSS team and Stu in New Zealand it could be very beneficial for all.
Nothing manevolent intended or assumed - just a different point of view. Each observer has their own set-up that will suit their way of doing things. One just needs to be careful when declaring something is better or is best that it may not suit all tastes.
In the work I do credit goes to the observer who's data first uncovered a probable binary detection and is assigned by the PI. This may not be the PO (principal observer). but everone who participates in obtaining data during the discovery apparition is credited in order of amount of contributed data. The PI always gets second credit. Each field will have its own rules. My comment was based on what appears, in some cases,(both amateurs and pros are guilty of this) to be more effort put into ensuring they get credit than ensuring the quality of the observations being made. Good science is good observing....
My own discoveries thus far have been a direct comparison with my reference images, no blinking just real time eyeballing. If you decide to blink the images it may open up other problems - if you leave them until next day to check then you may lose your discovery to others ie CHASE etc.
I'm not quite clear on the difference of what you're talking about here Peter, hoping you can explain a little more....
Presumably whether you are blinking or eyeballing the difference you have a comparison image and a new image, so why does it matter if you are blinking or looking at it side-by-side with respect to how long beforey ou have a comparison image?
The theory I was going on was that I would capture images and compare to my own images later in the future (obviously this is stemming from my building up of a library of NGC images).
What I do is spend much of my time in the observatory doing side by side comparisons mostly as well as some blinking, both in real time eyeballing each image. This gives me a reasonable advantage in that if I spot something that needs follow up I can take additional images straight away, not waiting until the next night (if I had checked the image several hours later). This comes at the expense of numbers and with SN hunting numbers are important. Building up a reference library is very important indeed, source it from your own images DSS images etc.
Following on from this thread, I now have my first night of specifically SN dedicated images. I got 128 objects last night, 3 images of each.
Blinking them is proving to be the time consuming and error prone bit for me so far.
Question: What do I blink my image against if the DSS image of the galaxy is too "contrasty" for me to be sure if the star I'm seeing is normal or not? I'm using THeSky6 so viewing the DSS image in there, having searched for the PGC object. Google Image search is my next stop, but for PGC's it rarely comes up with a suitable result.
For example in the attached I'm suspecting the star in my image is hidden in the bright area of the DSS image, but cannot be sure.
Blink against your own images. In the past I have taken my initial set, then each other night I take a single image of each target and blink against my initial set (this assumes the initial set did not contain SN).
ok, thanks. Obviously my problem here was I didn't have an inital set. However I've ruled out a SN by turning off galaxies in ThSky and seeing that there is a USNO star where my suspect is.
Hi Roger,
Here is one of my cropped images of PGC23496. You need to build up your own reference library of images asap. You have made a great start already. Then as you get better seeing replace the older images with better ones taken with the same equipment and exposures. This will help with the blinking. Use DSS images from http://archive.stsci.edu/cgi-bin/dss_form (the Sky 6 images are pretty much useless). I use the DSS Red and IR survey to check any suspects against as well. Astrometrica is fine for your ball park magnitude estimate at this stage and any discoveries are fine to report with this so long as you note your sources.
Alex Wassilieff from New Zealand did some great work on a CCD atlas and this is very useful. Also try looking at the DSS images as negatives, often stars near the nucleus may stand out more. We really need to talk on the phone, there is much to be gained by all , if you would like to please pm me with your number and I will call you tonight.
Blink against your own images. In the past I have taken my initial set, then each other night I take a single image of each target and blink against my initial set (this assumes the initial set did not contain SN).
Cheers
Taking single imges of targets, I can see why Peter likes checking them as they come down. Single image per target would get me 240 galaxies in 7 hours (current dark hours). If everyone works at that rate or faster (my system has things making it slow, so 240 strikes meas a low rate compared to what a more advanced setup would achieve), with even just a small number of people doing it, I can see how someone could easily beat you to it. Hmm.
I visual compare my images to a reference set of DSS images i had previously download. When the galaxy core of a DSS images is burnt out I then use one of my own.
I used some sofware called Aladin to bulk download 6000+ DSS images that I now use as my reference images.
I can image and check 40 images per hour, doing 60 second exposures.
Supernova discovery software
