Euclid NIR image simulation Gregor Seidel Max Planck Institute for - - PowerPoint PPT Presentation

▶
euclid nir image simulation
SMART_READER_LITE
LIVE PREVIEW

Euclid NIR image simulation Gregor Seidel Max Planck Institute for - - PowerPoint PPT Presentation

Jan 17, 2023 138 likes •507 views

Euclid Consortium Euclid NIR image simulation Gregor Seidel Max Planck Institute for Astronomy Heidelberg EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12 Euclid imagem Consortium what: (1)image / point source simulation

slide-1
SLIDE 1

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

Euclid NIR image simulation

Gregor Seidel Max Planck Institute for Astronomy Heidelberg

slide-2
SLIDE 2

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

imagem

  • what:

(1)image / point source simulation and testing pipeline (2)written in C++, Linux, command line user interface

  • dependencies:

dependencies: libfftw3-dev libcfitsio3-dev libpng12-dev libcairo2-dev libreadline6- dev libpstreams-dev libpthread-stubs0-dev

  • initial purpose:

determine limiting magnitudes for flat-spectrum point sources (1)gauge influence of individual instrument model parameters (2)optimise requirements on instrument performance margins

  • goal:

complete NIR imaging module for end to end simulation pipeline

slide-3
SLIDE 3

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

simulation pipeline

dithering strategy further dithers cosmic rays

slide-4
SLIDE 4

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

source image

dithering strategy further dithers cosmic rays

slide-5
SLIDE 5

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source image

Skylens+imagem: single H-band detector, using 3 dithers

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-6
SLIDE 6

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source image

Skylens+imagem: single H-band detector, using 3 dithers

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-7
SLIDE 7

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source image

Skylens+imagem: single H-band detector, using 3 dithers

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-8
SLIDE 8

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source image

Skylens+imagem: single H-band detector, using 3 dithers

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-9
SLIDE 9

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source image

Skylens+imagem: single H-band detector, using 3 dithers

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-10
SLIDE 10

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source

imagem: H band window 100x100 pixels Sérsic galaxy, realistic PSF from spectrum (Elisabete Da Cunha)

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-11
SLIDE 11

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source

imagem: Y band grid of 10 x 10 point sources convolved with PSF and background added

  • input fluxes from:

(1) Skylens (Massimo Meneghetti, Peter Melchior):

  • raytracing simulation

accounts for lensing

  • input from cosmological

simulations

  • Sérsic or HUDF galaxies
  • deconvolution through

wavelet decomposition (2) Sérsic models for galaxies, galaxy and star catalogues (3) Point source grids

slide-12
SLIDE 12

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

  • ptics & filter throughput

dithering strategy further dithers cosmic rays

slide-13
SLIDE 13

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

  • ptics & filter throughput
  • throughputs and spectral energy distribution to

(1) convert apparent magnitude to flux (2) superimpose wavelength dependent PSF (see below)

  • for arbitrary filters

(1) get, for any band, 5-sigma limiting magnitudes (2) adjust Y, J, H exposure times to equal limiting magnitude for all bands

slide-14
SLIDE 14

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

  • ptical PSF

dithering strategy further dithers cosmic rays

slide-15
SLIDE 15

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

  • detailed treatment of the optical PSF using PSF database (Frank Grupp,

Rory Holmes)

  • versampled PSFs for Y-, J-, H-band filters and 9 field positions from

920nm to 2000nm in ~14nm steps

  • ptical PSF

PSFs (shown at 1.4um, 1.6um, 2um) at field position 1, central 522x522 1μm2 sub-pixels, logarithmic scaling combined H-band PSF

...

slide-16
SLIDE 16

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

  • generation of synthetic PSFs of either

(1) double Gaussian or (2) ring shape

  • determination of EE50 and EE80 radii
  • signal-to-noise tests:
  • ptical PSF

...

  • point source S/N not

sensitive to PSF wings

  • Size of faint sources

comparable to pixel size

  • S/N driven by flux on

central pixel

slide-17
SLIDE 17

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

intrapixel response

dithering strategy further dithers cosmic rays

slide-18
SLIDE 18

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

intrapixel response

  • the quantum efficiency can vary on the

surface of each pixel

  • depending on the shape and strength of the

variation, the photometry then varies with the sub-pixel position of a source

  • result: intra-pixel variation < 5.5% to obtain

photometric error < 0.8% in 3 dithers measured intra-pixel response variations (N. Barron) 1-dither photometric precision for 5% intra-pixel response variation and varying FWHM of the AOCS + detector (not optical) PSF

0.8% error in 3 dithers 5% intra-pixel variation

slide-19
SLIDE 19

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

flat, dark & readnoise

dithering strategy further dithers cosmic rays

slide-20
SLIDE 20

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

flat, dark & readnoise

  • perability

(logscale)

  • varying quantum efficiency, dark

current and readout noise per pixel

  • can generate histograms from

existing detector maps or a given mean and standard deviation

  • can generate maps from given

histograms

  • flatfielding, dark current

subtraction and weight image for each dither; weights taken into account for drizzling

  • => operability criterium
slide-21
SLIDE 21

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

flat, dark & readnoise

single dither (logscale)

  • varying quantum efficiency, dark

current and readout noise per pixel

  • can generate histograms from

existing detector maps or a given mean and standard deviation

  • can generate maps from given

histograms

  • flatfielding, dark current

subtraction and weight image for each dither; weights taken into account for drizzling

  • => operability criterium
slide-22
SLIDE 22

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

flat, dark & readnoise

drizzled image (logscale)

  • varying quantum efficiency, dark

current and readout noise per pixel

  • can generate histograms from

existing detector maps or a given mean and standard deviation

  • can generate maps from given

histograms

  • flatfielding, dark current

subtraction and weight image for each dither; weights taken into account for drizzling

  • => operability criterium
slide-23
SLIDE 23

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

signal-to-noise test

dithering strategy further dithers cosmic rays

slide-24
SLIDE 24

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

signal-to-noise test

imagem: Y band grid of 10 x 10 point sources convolved with PSF and background added

  • input fluxes from point source

grids

slide-25
SLIDE 25

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

signal-to-noise test

imagem: Y band grid of 10 x 10 point sources convolved with PSF, background added and pixelised

  • input fluxes from point source

grids

slide-26
SLIDE 26

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

signal-to-noise test

imagem: Y band grid of 10 x 10 point sources convolved with PSF, background added, pixelised with quantum efficiency and flat, dark and readnoise maps applied imagem: the same, drizzled, flatfielded and dark subtracted

slide-27
SLIDE 27

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

signal-to-noise test

imagem: Y band grid of 10 x 10 point sources convolved with PSF, background added, pixelised with quantum efficiency and flat, dark and readnoise maps applied imagem: the same, drizzled, flatfielded and dark subtracted

slide-28
SLIDE 28

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

signal-to-noise test

imagem: Y band grid of 10 x 10 point sources convolved with PSF, background added, pixelised with quantum efficiency and flat, dark and readnoise maps applied imagem: the same, drizzled, flatfielded and dark subtracted

slide-29
SLIDE 29

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

  • assume:

a priori PSF knowledge ‑

  • centroiding:

fit drizzled PSF for every 18 x 18 sub-pixel position in 3 x 3 pixel square

  • extract photometry from

best fit to drizzled image

signal-to-noise test

slide-30
SLIDE 30

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

known unknowns

dithering strategy further dithers cosmic rays e- persistence non-linear effects ghost images diffraction spikes

slide-31
SLIDE 31

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source dither drizzled image

  • ptical PSF

AOCS PSF background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

known unknowns

dithering strategy further dithers cosmic rays e- persistence non-linear effects ghost images diffraction spikes

slide-32
SLIDE 32

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

performance test

  • Poisson distributed background
  • single band
  • 2040 x 2040 pixel detector
  • 18 x 18 sub-pixels per pixel
  • realistic flat, dark and readout-noise maps
  • single CPU processing time on 2.80GHz Intel Xeon X5560:

13min 46s

  • maximum RAM usage: ~21GB (double precision)
slide-33
SLIDE 33

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

survey requirements estimate

  • assuming:

(1)15000 square degrees and 0.3” pixels (2)~50% in 3 dithers (3)~50% in 4 dithers (4)2 bytes per pixel (5)no overlap, no deep survey, no flats / darks / etc.

  • ~14TB wide survey storage requirement (FITS)
  • ~420k CPU hours
slide-34
SLIDE 34

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

simulation pipeline 2.0

  • next step: sky simulation
slide-35
SLIDE 35

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

source shape+spectrum dither drizzled image

  • ptical+AOCS PSF

background + ∗ incident flux pixelised flux system response intra-pixel response electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues ⋅ pixel size

simulation pipeline 2.0

dithering strategy further dithers cosmic rays

...

source shape+spectrum

  • ptical+AOCS PSF

∗ incident flux pixelised flux system response intra-pixel response ⋅ pixel size

slide-36
SLIDE 36

EUCLID CSWG + OUSIM meeting Barcelona Gregor Seidel 10/07/12

Euclid Consortium

dither drizzled image background + electron image exposure time flat-field dark-map Poisson noise readout-noise pixel crosstalk weight maps e-/ADU conversion S/N tests data analysis pipelines catalogues

simulation pipeline 2.0

dithering strategy further dithers cosmic rays stamp-1 transformation (ra,dec,z)→(detector #,x,y)

  • ptical distortions

detector placement pointing stamp-42 stamp-3 stamp-4 stamp-2

...

stamp-N

...

deflection map / lensing focal plane position distortion sub-pixel placement