LOFAR TKP meeting Amsterdam, December 4th 2012 Source Finder Testing Overview & Status Dario Carbone 

LOFAR Source Finders WG Lead by Chiara Ferrari R. Breton, D. Carbone, P. Carroll, A. Dabbech, H. Garsden, A. van der Horst, A. Mints, R. Paladino, D. Rafferty, A. Rowlinson, A. Shulevski, J. Swinbank, S. van Velzen. http://www.lofar.org/wiki/doku.php?id=commissioning:source_finders 

LOFAR Source Finders WG Lead by Chiara Ferrari R. Breton, D. Carbone, P. Carroll, A. Dabbech, H. Garsden, A. van der Horst, A. Mints, R. Paladino, D. Rafferty, A. Rowlinson, A. Shulevski, J. Swinbank, S. van Velzen. http://www.lofar.org/wiki/doku.php?id=commissioning:source_finders Main aims ● Identify the best settings for automatically running source finders in LOFAR/MSSS/TRAP pipelines ● Report bugs and needs to developers and provide support to LOFAR users ● LOFAR representative in the international radio source finding discussion 

Work done & to be done DONE ● Tests on real LOFAR data ● Presence of both testers and developers ● Finding & fixing bugs ● Evaluate the performance of the SF tools 

Work done & to be done DONE ● Tests on real LOFAR data ● Presence of both testers and developers ● Finding & fixing bugs ● Evaluate the performance of the SF tools TO BE DONE ● Test SF tools on reliable simulated LOFAR maps ● Participate to the international campaign on SF by Hancock et al. 

Tests & users assistance WENSS map by R. Pizzo 

Tests & users assistance Test and development by D. Rafferty and C. Ferrari 

Tests & users assistance LOFAR map of GRS1915 by D. Carbone Source extraction by D. Rafferty 

Systematic position errors ● Very important for the transient pipeline ● Source could not to be associated ● Tests on 70 images of 10 fields, with a pixel size of 45 arcsec. 

Systematic position errors ● ● PyBDSM loses 13.3% PySE 3.4% R DR = 5.68 Credits to A. Van der Horst 

Systematic position errors ● Systematic position error of 20 arcsec ● PyBDSM loses 2.0% PySE 1.0% Credits to A. Van der Horst 

Flux errors Circles: objects both in the VLSS and the LOFAR catalogs (extracted with PyBDSM and/or PySE) Differences : < 40% < 70% < 80% > 80% Credits to C. Ferrari 

Statistics on source detection 

Diagnostic of non detection 

Questions at SF WG meeting i. Which source finders should we keep testing? We should focus our tests on PySE and PyBDSM. 

Questions at SF WG meeting ii. Which are the parameters that are important to set for optimizing source finders? Detection threshold, Analysis threshold, False Detection Rate, RMS box and Adaptive RMS box, Margins or radius, Beam infos, De-blending and Extended sources detection i. Which source finders should we keep testing? We should focus our tests on PySE and PyBDSM. 

Questions at SF WG meeting iii. Which corrections should we introduce in theoretical error bars given by the source finder tools? The position errors are underestimated by PyBDSM and systematics errors have to be introduced. Better constrains will be put using simulated maps. ii. Which are the parameters that are important to set for optimizing source finders? Detection threshold, Analysis threshold, False Detection Rate, RMS box and Adaptive RMS box, Margins or radius, Beam infos, De-blending and Extended sources detection i. Which source finders should we keep testing? We should focus our tests on PySE and PyBDSM. 

Questions at SF WG meeting iv. Which source association tools and strategy should we use in order to combine different catalogs? B. Scheers's method: the De Ruiter radius, taking into account position errors and thus preventing spurious source associations. CAVEAT: without appropriate systematic position errors very bright sources could be NOT associated. 

Questions at SF WG meeting iv. Which source association tools and strategy should we use in order to combine different catalogs? B. Scheers's method: the De Ruiter radius, taking into account position errors and thus preventing spurious source associations. CAVEAT: without appropriate systematic position errors very bright sources could be NOT associated. v. Which source extraction and association strategy should we use for multi-band LOFAR catalogs? MSSS scripts and spectral module of PyBDSM being tested to make multi-band LOFAR catalogs. Could be implemented in PYSE if desirable. 

Simulated maps ● Procedure: – input Gaussian noise in the visibilities of a true LOFAR MS to have the directional dependent gains already in place. – input simulated sources in the visibilities, centered on the pointing center of the original LOFAR MS, keeping the instrument table untouched. The simulated source must be in BBS format. – calibrate and image the obtained MS as a normal observation. 

Simulated maps Simulated LOFAR map by A. Rowlinson 

Simulated maps Simulated LOFAR map by A. Rowlinson A very bright source has been added 

Conclusion Main aims of LOFAR SF WG: ● Identify the best settings for automatically running source finders ● Report bugs to developers and provide support to LOFAR users Work done: ● Tests on real LOFAR data ● Evaluate the performance of the SF tools Work to be done: ● Test SF tools on reliable simulated LOFAR maps 

SF WG wikipage http://www.lofar.org/wiki/doku.php?id=commissioning:source_finders 

No content