Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.atnf.csiro.au/research/workshops/2010/sourcefinding/Hancock.pdf Дата изменения: Mon May 10 06:33:07 2010 Дата индексирования: Sun Jun 27 15:01:03 2010 Кодировка: Поисковые слова: столовая гора

The Future of Continuum Source Finding for ASKAP SSPs

Paul Hancock Sydney Institute for Astronomy (SIfA)


Source Finding today

> Generic SF programs are available
- Tend to be quite basic
- Require post-processing

> More refined SF programs
- Telescope/Survey specific
- Effective - Hard to transfer to other Telescopes/Surveys

> ASKAP will require a refined program > "How well do the currently available source finding programs work on the simulated ASKAP data?"


What do we want from a SF program?

> Completeness
- As many real source as possible

> Reliability
- As few spurious sources as possible

> A predictable behaviour of the above
- To statistically correct for the effects of completeness and reliability - e.g., how does the reliability change with SNR or flux density?

> Automation is essential
- Computers are for data crunching - Humans are for doing science

> Fast / cpu efficient


What DATA do we want from a SF program?

> Position/Flux + err

> Morphology
> Local noise level > Detection confidence

> Complete or filtered source list

4


Evaluate Current SF Methods

Total Intensity > SUMSS/MGPS (IMSAD)

Polarized Intensity > VSAD

> S-extractor
> Duchamp > sfind > + Home-Brew Programs

>?


The AT20G source detection algorithm
Inject Sources
> Known list of sources

- MC simulation of the source distribution
> Real background data

- Noise
- Instrumental problems - Coverage variations

- Missing data


The AT20G source detection algorithm
Recover and Compare
> Find sources

> Record quality parameters
> Compare with auxiliary data > Separate Real/Spurious sources > Measure recovery rate, reliability and completeness as a function of flux/position/SNR/etc

7


ASKAP application

> ASKAP Simulations (Wilman et al. 2008, Obreschkow et al 2009)
- Sky model (sources) - Noise image - Restored image

> Enough information is available to test source finding programs

8


Future Plans

> SF will have high priority

> I will lead/coordinate the source finding program testing
> Testing of your own / favourite program is welcome
- Using the SF testing to hone your program is not cheating!

> Source finding programs to be tested on ASKAP and ATLAS
> Possible Outcomes: 1)Small amount of work to be done on current SF program(s) 2)Large amount of work to be done on current SF program(s) 3)Completely new program is required 4)Best SF is project/science dependant (don't think so)
9


Tell me what you want from a SF program

> High completeness

> Good reliability
> ...

10