1
ModernObservational/Instrumentation
TechniquesAstronomy 500
Andy Sheinis, Sterling 5520,[email protected] 2:30, 6515 SterlingOffice Hours: Tu 11-12
• Ch1 of Birney, Gonzalez and Oesperhas coordinate stuff
2
Preliminary Processing
• Image display• Bias subtraction• Flat fielding• Photometric Calibration• Sky subtraction• Wavelength calibration• Spectro-photometric calibration
Preliminary Processing
• There are two types of instrumentalsignature to remove:– Additive:
• Bias Level• Bias Structure• Dark Counts
– Multiplicative:• Q.E. variations on all scales
Constant # of countsadded independent ofthe brightness of thesource(s).
Constant fractionaleffect
3
Astronomy Packages
• IRAF (http://iraf.noao.edu)• IDL (astronomy users library:
http://idlastro.gsfc.nasa.gov/homepage.html)• XVISTA (http://astro.nmsu.edu/holtz/xvista)• FIGARO• MIDAS• AIPS• Add-on packages: DAOPHOT
Astronomy Packages
Issues: availability, cost, philosophy, data handling (disk vs. memory), speed, ease of use (e.g., keywords vs. parm files), language and access to existing code, ability to add new code, scripts/ procedures (internal control language).
4
Image Display
• DS9 or SAO image with IRAF• ATV with IDL• Issues:
– Dynamic Range (8-bit)– Sampling (spatial and depth)– Scaling (depth)– Color map
Image file formats
• FITS (Flexible Image Transport System)• Header (80 byte ASCII records)• Data (16-but Int, 32 bit int or 32 bit FP)• Details of the data are conmtained in the
header keywords, I.e.– Simple– Bitpix– Naxis– Ad infinitum
5
Image file formats
• IRAF uses internally defined data format(BAD!)
• Used for all (disk based) calculations• Proprietary and subject to change!• As an option, they now offer FITS format
calculations. (I highly recommend this!)• Not the default, need to add something to
your login.cl file
Preliminary Processing
• Bias subtraction– Model bias level– Subtract (may be different for different
readout amplifiers)– Definitely different for mosaiced images
6
Bias Correction
• Bias level and any y (along columns)gradient is taken out via overscansubtraction.
• Bias structure is taken out bysubtracting a zero-level frame.
• In IRAF ccdproc takes care of both.
Overscan
• After reading out theCCD `real’ pixels, youcan continue to read outvirtual pixels and recordthe bias level and readnoise of the amplifiers.These virtual pixels arecalled the overscanregion
Active area of CCD
Row orline #
Column # overscan region
7
Two amplifier overscan
8
Overscan subtraction
• First you need toidentify therelevantcolumns.
• In IRAF use theformat:
[x1,x2:y1,y2]
Sky level
Overscan#1Oscan#2
In IRAF, plots like this are made using implot
Preliminary Processing
• Flat fielding– 2 kinds high order, pixel-to-pixel variations– Low order, sky illumination, vignetting,
scattered light– Model the flat field based on dome flats,
sky flats, twilight flats– Divide by normalized flat field
9
Flat Fielding
• There are pixel-to-pixel QE variations and lowerspatial frequency QE variations in all electronicdetectors. The goal of flat-fielding is to multiply everypixel by the correct normalizing factor to eliminatethese QE differences.
• Ideally, illuminate the detector with a source that is asflat on the sky as the background and collect at leasta million e- per pixel. Then could flat-field to=0.1%.
!
106
106
= 0.001
Flat Fielding
• If you could illuminate the CCD uniformly,then normalize the mean to 1, this imagecould be divided into every frame.
• For direct imaging, usually use a combinationof:– Dome Flats– Twilight Flats– Dark Sky Flats
10
Dome Flats
• Put some quartz (hot, continuum source) lamps onthe telescope and illuminate a white screen or spoton the dome.
• These often don’t work very well for two reasons:– The lamps are always too cool (red)– The dome is not even close to infinity and usually illuminates
the primary differently than the sky
• But, you can collect a lot of photons during the day
Twilight Flats
• These often work pretty well• The Sun is pretty hot, the scattering
surface illuminates the telescope justlike the dark night sky
• Doesn’t use dark time
11
Dark-sky Flats
• These tend to work very well. Theymatch the sky perfectly
• They sometimes require useful darktime
• They sometimes contain fringes
Stars and Galaxies
• For twilight and dark sky flats you have a problem inthat they contain stars and galaxies. The usual trick isto move the telescope between exposures and thendo a non-registered stack of the frames in each filter.
• Median or better yet minmax rejection (for example,in the frame combining can effectively eliminate allthe stars and galaxies in the combined flat.
12
Minmax rejection
Pixel value: 1000 200 180 180
Average with minmax rejection, reject 2 highest value averaginglowest two will give the sky value. NOTE! Must normalizeframes to common mean or mode before combining! Sometimesit is necessary to pre-clean the frames before combining.
Combining Frames
• In IRAF, imcombine is the task tocombine frames.– combine=average– reject=minmax– scale=mode– nlow=0– nhigh=2
13
Combining Frames
• In IDL, just do arithmetic on the images• Image C=Image A+Image B• Can use canned routines such as:
“medarr,bigarray,outarray”
Flat fielding tricks• Combine domes (high counts, bad illumination) with
dark sky flats (low counts, excellent illumination).1. Spatially smooth (or fit low-order surface to) both combined
dome and combined dark sky. > sDome, sDark2. Remove dome low-spatial-frequency pattern:
Dome/sDome3. sDark is the sky flat with the low-f pattern already removed.4. Best of both worlds is (Dome/sDome) x sDark
14
PFCam flat fieldsI-band
V-band
U-band
Dust on filter
Rings due to non uniform thinning
Note differences with color.This means that objects withdifferent spectra will be flat-fielded slightly incorrectly.
Two amplifer readout
15
Flat-field tests
1. Take cuts through your flat-fielded framesand make sure the sky is flat (checkcorners). IRAF implot, IDL (ATV (r and c)implot)
2. In blank areas, make sure the pixel-to-pixelvariations are consistent with shot noisefrom the sky level. IRAF imexam and the`m’ key, IDL (ATV, imexam, photometry)
Preliminary Processing
• Photometric Calibration– Measure R* for real, using a known
standard through the exact same setupand conditions
– Find a nearby standard– May want to observe before and after
16
Preliminary Processing
• Wavelength calibration (forspectroscopy
• Measure arcd lamps, and/or night skylines
• Fit lines to a model• Check and modify
Preliminary Processing
• Spectro-photometric calibration(spectroscopy)
• Measure R* for a photometric standard underthe same conditions
• Get photon caount at every wavelength bin