bernhard hubl cedic 2009 colour calibration of rgb images advantages and disadvantages of different...
TRANSCRIPT
Bernhard Hubl CEDIC 2009
Colour Calibration of RGB Images
Advantages and disadvantages of different colour calibration methodes
Contents
Field Stars Calibration G2 Star Calibration
• Procedure• Disadvantages
B-V Calibration• Workflow• Experiences
Alternative Methodes Summary
White Balance
Average of field stars = white Sunlight = white
• Stars with sun-like spectral type (G2)• Stars with sun-like photometry (B-V)
Contents
Field Stars Calibration G2 Star CalibrationG2 Star Calibration
• ProcedureProcedure• DisadvantagesDisadvantages
B-V CalibrationB-V Calibration• WorkflowWorkflow• ExperiencesExperiences
Alternative MethodesAlternative Methodes SummarySummary
Field Stars Calibration
Assumption• Average colour of bright field stars is white
Advantages• Easy (e.g. AIP4Win)• Independent of atmospheric extinction• Works fine for galaxy fields
Disadvantages• Open star clusters• Interstellar extinction• Bright nebulae
Contents
Field Stars CalibrationField Stars Calibration G2 Star Calibration
• ProcedureProcedure• DisadvantagesDisadvantages
B-V CalibrationB-V Calibration• WorkflowWorkflow• ExperiencesExperiences
Alternative MethodesAlternative Methodes SummarySummary
Workflow G2 Star Calibration
Measurement of bright G2 star
RGB weightsin zenith
Exposure of object in specific altitude
Calculate extinction
Corrected RGB-
weights
When does the G2 calibration fail?
No measurement of a G2 star Very low altitude of the object Bad transparency (clouds, fog,..)
Measurement of bright G2 star
RGB weightsin zenith
Exposure of object in specific altitude
Calculate extinction
Corrected RGB-
weights
Object exposures for G2 star calibration
Idea: Search for not saturated G2 stars Two problems:
• Spectral type of faint stars is often unknown• Reddening of G2 strars by interstellar dust.
Contents
Field Stars CalibrationField Stars Calibration G2 Star CalibrationG2 Star Calibration
• ProcedureProcedure• DisadvantagesDisadvantages
B-V Calibration• WorkflowWorkflow• ExperiencesExperiences
Alternative MethodesAlternative Methodes SummarySummary
B-V Calibration
Idea: Peter Riepe & Harald Tomsik• VdS Journal 25+26
Calibration uses the object image Instead of spectral type -> photometry
• Comprehensive photometric data of faint stars• Johnson-Cousins UBVRI filter system.
Colour Index
= difference of the brightness measured through two different filters
e.g: B-V, V-R Sp. type Colour B-V [mag]
Spica B1 Blue -0.23
Deneb A2 Blue-White +0.09
Sun G2 White +0.65
Betelgeuse M0 Orange-Red +1.50
Idea
Colour index of a G2 star without interstellar extinction:• B-V=0.65 mag• V-R=0.36 mag
Search for white stars inside the image field:• Not saturated• 0.6 mag < B-V < 0.7 mag• 0.2 mag < V-R < 0.6 mag
Measurement of the brightness of selected white stars in Red, Green and Blue.
B-V Workflow
Preprocessing• Dark, Flat• Registration• Averaged R, G and B image
Aladin B-V steps1. Load DSS image2. Load photometric catalogue3. Load filter4. Measure brightness of selected stars
Step2: Load Photometric Catalogue
NOMAD1 = Naval Observatory Merged Astrometric DatasetAdvantage: 100% sky coverageDisadvantage: Accuracy
Step3: Load Filter
•Select white stars•0.6 < B-V < 0.7•0.2 < V-R < 0.6
•Save Wolfgang Piracher‘s filter on your local disk
Step4: Measurement
Aperture photometry of standard image processing software (e.g. AstroArt)
Result: 3 ADU values (R,G and B) Example:
• Red 48298 / Green 56917 / Blue 51635• -> Green and Blue are too bright• -> Red 1,00 / Green 0,85 / Blue 0,94.
Experiences with B-V
Methode is laborious NOMAD1 catalogue
• Inhomogeneous• Is based partly on plate scans• The fainter the reference stars, the more unreliable
-> at least 5 stars are necessary (outliers) The blue weight is often too high Average between G2 and B-V.
Contents
Field Stars CalibrationField Stars Calibration G2 Star CalibrationG2 Star Calibration
• ProcedureProcedure• DisadvantagesDisadvantages
B-V CalibrationB-V Calibration• WorkflowWorkflow• ExperiencesExperiences
Alternative Methodes SummarySummary
Alternative Methodes
Regim by Andreas Rörig• Up to 5 stars; manual selection
Theli by Mischa Schirmer• LINUX software• Astrometric registration• Fully automated B-V methode: Usage of a big number of
stars, which are found by the software (Vizier data base).
SDSS• Much better photometric data• Filter system ugriz (u-g=1.43mag; g-r=0.44mag)• At the moment far away from full sky coverage.
Summary
New catalogues• In several years much better catalogues compared with
NOMAD1 Fully automated B-V methode is possible
• E.g. Theli B-V methode as standard
• The B-V calibration works in all thinkable situations (bad transparency, Milky Way fields,..), provided that the used catalogue is enough accurate.
Best procedure at the moment:• Combination of G2 and B-V