IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links

Changeset 39850 for trunk


Ignore:
Timestamp:
Dec 12, 2016, 6:03:54 PM (10 years ago)
Author:
watersc1
Message:

'Final' draft edition. Added diff discussion, removed majority of remaining red text, with the exception of the unique object count, which I'm not sure of. References are still missing and need to be resolved.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/doc/release.2015/ps1.detrend/detrend.tex

    r39844 r39850  
    4848% list and (2) re-order the list at the bottom (and comment-out as needed)
    4949\def\IfA{1}
    50 \def\CfA{2}
    51 \def\MPIA{3}
    52 \def\Princeton{3}
    53 \def\USNO{4}
    54 \def\JHU{1}
     50\def\Princeton{2}
     51\def\STSCI{3}
     52\def\Pitt{4}
     53%\def\CfA{2}
     54%\def\MPIA{3}
     55%\def\USNO{4}
     56%\def\JHU{1}
    5557
    5658% This example has a first author from UH:
    5759\author{
    58 C. Z. Waters,\altaffilmark{\IfA}
    59 IPP Team,
     60C.~Z. Waters,\altaffilmark{\IfA}
     61E.~A. Magnier,\altaffilmark{\IfA}
     62P.~A. Price,\altaffilmark{\Princeton}
     63H.~A. Flewelling,\altaffilmark{\IfA}
     64M.~E. Huber,\altaffilmark{\IfA}
     65W.~E. Sweeney,\altaffilmark{\IfA}
     66J.~L. Tonry, \altaffilmark{\IfA}
     67K.~C. Chambers,\altaffilmark{\IfA}
     68R.~H. Lupton,\altaffilmark{\Princeton}
     69A. Rest,\altaffilmark{\STSCI}
     70W.~M. Wood-Vasey,\altaffilmark{\Pitt}
     71PS1 Builders
    6072%PS Builder List
    6173% W.~S. Burgett,\altaffilmark{\IfA}
    62 % K.~C. Chambers,\altaffilmark{\IfA}
    6374% L. Denneau,\altaffilmark{\IfA}
    6475% P. Draper,\altaffilmark{\DUR}
    65 % H.~A. Flewelling,\altaffilmark{\IfA}
    6676% T. Grav,\altaffilmark{\IfA}
    6777% J. N. Heasley,\altaffilmark{\IfA}
    6878% K. W. Hodapp,\altaffilmark{\IfA}
    69 % M. E. Huber,\altaffilmark{\IfA}
    7079% R. Jedicke,\altaffilmark{\IfA}
    7180% N. Kaiser,\altaffilmark{\IfA}
    7281% R.-P. Kudritzki,\altaffilmark{\IfA}
    7382% G. A. Luppino,\altaffilmark{\IfA}
    74 % R. H. Lupton,\altaffilmark{\Princeton}
    75 % E. A. Magnier,\altaffilmark{\IfA}
    7683% N. Metcalfe,\altaffilmark{\DUH}
    7784% D. G. Monet,\altaffilmark{\USNO}
    7885% J.~S. Morgan,\altaffilmark{\IfA}
    7986% P. M. Onaka,\altaffilmark{\IfA}
    80 % P.~A. Price,\altaffilmark{\Princeton}
    8187% C.~W. Stubbs,\altaffilmark{\CfA}
    82 % W.~E. Sweeney,\altaffilmark{\IfA}
    83 % J.~L. Tonry, \altaffilmark{\IfA}
    8488% R. J. Wainscoat,\altaffilmark{\IfA} and
    8589% C. Z. Waters,\altaffilmark{\IfA}
     
    8993\altaffiltext{\IfA}{Institute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu HI 96822}
    9094% \altaffiltext{\CfA}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138}
    91 % \altaffiltext{\Princeton}{Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544, USA}
     95\altaffiltext{\Princeton}{Department of Astrophysical Sciences, Princeton University, Princeton, NJ 08544, USA}
     96\altaffiltext{\STSCI}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218, USA}
     97\altaffiltext{\Pitt}{Department of Physics and Astronomy, University of Pittsburgh, Pittsburgh, PA 15260, USA}
    9298% \altaffiltext{\USNO}{US Naval Observatory, Flagstaff Station, Flagstaff, AZ 86001, USA}
    9399% \altaffiltext{\JHU}{Department of Physics and Astronomy, Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218, USA}
     
    95101\begin{abstract}
    96102
    97 Lorem ipsum dolor sit amet, consectetur adipiscing elit. Vestibulum
    98 bibendum nisi id tristique posuere. Duis eu mollis nulla. Maecenas est
    99 turpis, mattis tempor urna vitae, placerat rhoncus sem. Lorem ipsum
    100 dolor sit amet, consectetur adipiscing elit. Sed quis velit
    101 nisl. Aliquam erat volutpat. Cras lacinia, nisl tristique auctor
    102 molestie, dolor nulla rhoncus purus, ac accumsan nunc nunc ac
    103 nibh. Maecenas vitae mollis mauris. Ut sollicitudin pulvinar purus,
    104 eget luctus lorem tincidunt vitae. Vestibulum eu mattis neque. Nulla
    105 in tortor id urna dapibus gravida a vel leo.
     103the Pan-STARRS1 Science Consortium have carried out a set of imaging
     104surveys using the 1.4 giga-pixel GPC1 camera on the PS1 telescope.  As
     105this camera is composed of many individual electronic readouts, and
     106covers a very large field of view, great care was taken to ensure that
     107the many instrumental effects were corrected to produce the most
     108uniform detector response possible.  We present the image detrending
     109steps used as part of the processing of the data contained within the
     110public release of the Pan-STARRS1 Data Release 1 (DR1).  In addition
     111to the single image processing, the methods used to transform the
     112375,573 individual exposures into a common sky-oriented grid are
     113discussed, as well as those used to produce both the image stack and
     114difference combination products.
    106115
    107116\end{abstract}
     
    131140Pan-STARRS 1 Science Consortium members.
    132141
    133 \czwdraft{Nigel: you mention calibrating to the reference catalog without telling us
    134 what this is composed of (maybe this is in a different section, but would be
    135 nice to have some idea here).}
    136 
    137 \czwdraft{Can we get around this point by simply adding a reference to
    138   the paper describing the reference catalog?  It's not really part of
    139   the detrending process, and is discussed here mostly to give an
    140   overview of the stages, and later to find sources of ghosts for
    141   masking.}
     142%% \czwdraft{Nigel: you mention calibrating to the reference catalog without telling us
     143%% what this is composed of (maybe this is in a different section, but would be
     144%% nice to have some idea here).}
     145
     146%% \czwdraft{Can we get around this point by simply adding a reference to
     147%%   the paper describing the reference catalog?  It's not really part of
     148%%   the detrending process, and is discussed here mostly to give an
     149%%   overview of the stages, and later to find sources of ghosts for
     150%%   masking.}
    142151
    143152The Pan-STARRS image processing pipeline (IPP) is described elsewhere
     
    167176identified in the \ippstage{diff} stage, which takes input
    168177\ippstage{warp} and/or \ippstage{stack} data and performs image
    169 differencing \citep{HuberXXX}.  Further photometry is performed in the
    170 \ippstage{staticsky} and \ippstage{skycal} stages, which add extended
    171 source fitting to the point source photometry of objects detected in
    172 the \ippstage{stack} images, and calibrate the results against the
    173 reference catalog.  The \ippstage{fullforce} stage takes the catalog
    174 output of the \ippstage{skycal} stage, and uses the objects detected
    175 in that to perform forced photometry on the individual \ippstage{warp}
    176 stage images.  The details of these stages are provided in
    177 \citet{MagnierXXY}.
     178differencing (Section \ref{sec:diffs}).  Further photometry is
     179performed in the \ippstage{staticsky} and \ippstage{skycal} stages,
     180which add extended source fitting to the point source photometry of
     181objects detected in the \ippstage{stack} images, and calibrate the
     182results against the reference catalog.  The \ippstage{fullforce} stage
     183takes the catalog output of the \ippstage{skycal} stage, and uses the
     184objects detected in that to perform forced photometry on the
     185individual \ippstage{warp} stage images.  The details of these stages
     186are provided in \citet{MagnierXXY}.
    178187
    179188The same reduction procedure described above is also performed in real
     
    192201details of the construction of those detrends in Section
    193202\ref{sec:detrend construction}.  An analysis of the algorithms used to
    194 complete the \ippstage{warp} (section \ref{sec:warping}) and
    195 \ippstage{stack} (section \ref{sec:stacking}) stage transformations of
    196 the image data to from the detector frame to a common sky frame, and
    197 the co-adding of those common sky frame images continues after the
    198 list of detrend steps.  Finally, a discussion of the remaining issues
    199 and possible future improvements is presented in section
    200 \ref{sec:discussion}.
     203complete the \ippstage{warp} (section \ref{sec:warping}),
     204\ippstage{stack} (section \ref{sec:stacking}), and \ippstage{diff}
     205(section \ref{sec:diffs}) stage transformations of the image data to
     206from the detector frame to a common sky frame, and the co-adding of
     207those common sky frame images continues after the list of detrend
     208steps.  Finally, a discussion of the remaining issues and possible
     209future improvements is presented in section \ref{sec:discussion}.
    201210
    202211Image products presented in figures have been
     
    218227and pixel $(1,1)$ to the top right of their position.
    219228
     229% Note taken verbatim from Ken's Paper 1.
     230\textit{Note: These papera are being placed on the arXiv.org to
     231  provide crucial support information at the time of the public
     232  release of Data Release 1 (DR1).  We expect the arXiv versions to be
     233  updated prior to submission and there could be significant
     234  variations with the refereed papers.  We apologize for the
     235  inconvience.}
     236
    220237% Discuss 2-phase/3-phase device differnces
    221238
     
    227244\label{sec:detrending}
    228245
    229 \czwdraft{Nigel: I forgot: when we are talking about the various bias corrections it might be
    230 worth pointing out that we expect these to be more of an issue in the g-band
    231 (and maybe r?) where read noise is a significant contributor.
    232 }
     246%% \czwdraft{Nigel: I forgot: when we are talking about the various bias corrections it might be
     247%% worth pointing out that we expect these to be more of an issue in the g-band
     248%% (and maybe r?) where read noise is a significant contributor.
     249%% }
    233250
    234251Ensuring a consistent and uniform detector response across the
     
    727744better represents the true detector response.
    728745
    729 \czwdraft{EAM: the flat-field construction part needs to make a clearer discussion of
    730 the skyflat vs the photometric correction (photflat) built initially for
    731 the survey vs the flat-field corrections determined in the database as part
    732 of ubercal (for the latter, you should just mention the concept -- it will
    733 also be mentioned in the calibration paper).  The statement that the
    734 flat-field response was stable is not true since we did need 5 'seasons'.}
     746%% \czwdraft{EAM: the flat-field construction part needs to make a clearer discussion of
     747%% the skyflat vs the photometric correction (photflat) built initially for
     748%% the survey vs the flat-field corrections determined in the database as part
     749%% of ubercal (for the latter, you should just mention the concept -- it will
     750%% also be mentioned in the calibration paper).  The statement that the
     751%% flat-field response was stable is not true since we did need 5 'seasons'.}
    735752
    736753In addition to this flat field applied to the individual images, the
     
    740757survey, five separate ``seasons'' of database flat fields were needed
    741758to ensure proper calibration.  This indicates that the flat field
    742 response is not completely fixed in time.
     759response is not completely fixed in time.  More details on this
     760process are contained in \citet{calibration}.
    743761
    744762\subsection{Pattern correction}
     
    954972    \includegraphics[width=1.5\hsize,angle=0,clip]{images/o5220g0025o_XY53_fringe.png}
    955973  \end{minipage}
    956   \caption{Example of the \yps{} filter fringe pattern on exposure o5220g0025o OTA53 (\yps{} filter 30s).  The left panel shows the OTA mosaic with all detrending except the fringe correction, while the right shows the same including the fringe correction.  Both images have been smoothed with a Gaussian with $\sigma = 3$ pixels to highlight the faint and large scale fringe patterns. \czwdraft{See if there's a way to have mana produce images larger than the screen size.}}
     974  \caption{Example of the \yps{} filter fringe pattern on exposure o5220g0025o OTA53 (\yps{} filter 30s).  The left panel shows the OTA mosaic with all detrending except the fringe correction, while the right shows the same including the fringe correction.  Both images have been smoothed with a Gaussian with $\sigma = 3$ pixels to highlight the faint and large scale fringe patterns.
     975%\czwdraft{See if there's a way to have mana produce images larger than the screen size.}
     976}
    957977  \label{fig: fringe example}
    958978\end{figure}
     
    13721392\label{sec:background}
    13731393
    1374 \czwdraft{Nigel: 2.10 The background section is rather short, given all the fuss DRAVG made
    1375 about it. What is done to eliminate contamination by bright objects - isn't
    1376 there some sort of clipping? We also have a confusing number of ``bins'' in the
    1377 text (``These bins have 10000 .... a binned cumulative distribution is
    1378 generated. These bins are interpolated ... levels. Repeating this across all
    1379 bins ...''). There is no mention of the fact that this will subtract real
    1380 astrophysics backgrounds if they are on a suitably large scale, or of the fact
    1381 that the subtraction is not perfect (don't I remember that the stacks end up
    1382 with a non-zero background which scales with the number of input warps?).
    1383 }
    1384 
    1385 \czwdraft{Based on the wiki page on 2014-05-21 the stack background issue should be resolved.}
     1394%% \czwdraft{Nigel: 2.10 The background section is rather short, given all the fuss DRAVG made
     1395%% about it. What is done to eliminate contamination by bright objects - isn't
     1396%% there some sort of clipping? We also have a confusing number of ``bins'' in the
     1397%% text (``These bins have 10000 .... a binned cumulative distribution is
     1398%% generated. These bins are interpolated ... levels. Repeating this across all
     1399%% bins ...''). There is no mention of the fact that this will subtract real
     1400%% astrophysics backgrounds if they are on a suitably large scale, or of the fact
     1401%% that the subtraction is not perfect (don't I remember that the stacks end up
     1402%% with a non-zero background which scales with the number of input warps?).
     1403%% }
     1404
     1405%% \czwdraft{Based on the wiki page on 2014-05-21 the stack background issue should be resolved.}
    13861406
    13871407Once all other detrending is done, the pixels from each cell are
     
    18131833assumed to be included in the zeropoint and transparency values.
    18141834
    1815 
    1816 \czwdraft{Nigel: 5. ``The ouput exposure time is set to the sum of the input exposure times.''
    1817 True, but we should note that as warps can be rejected later on in the
    1818 stacking process this output time is notional in some sense.
    1819 Calibration - for PV3 what photometric calibration has been used at this stage
    1820 for the input warps? Should we make it clear here that pixels are not subject
    1821 to the final (any?) ubercal?
    1822 }
     1835The zeropoint calibration performed here uses the calibration of the
     1836individual input exposures against the reference catalog.  Upon the
     1837conclusion of the survey, the entire set of detection catalogs is
     1838further re-calibrated in the ``ubercal'' process \citep{ubercal}.
     1839This produces a more consistent calibration of each exposure across
     1840the entire region of the sky imaged.  This further calibration is not
     1841available at the time of stacking, and so there may be small residuals
     1842in the transparency values as a result of this \citet{calibration}.
     1843
     1844%% \czwdraft{Nigel: 5. ``The ouput exposure time is set to the sum of the input exposure times.''
     1845%% True, but we should note that as warps can be rejected later on in the
     1846%% stacking process this output time is notional in some sense.
     1847%% Calibration - for PV3 what photometric calibration has been used at this stage
     1848%% for the input warps? Should we make it clear here that pixels are not subject
     1849%% to the final (any?) ubercal?
     1850%% }
    18231851
    18241852% PREPARE
     
    20412069
    20422070\begin{eqnarray}
    2043   \mathrm{limit}_\mathrm{mixture model} &=& 4^2 * (\sigma^2_\mathrm{input} + \sigma_\mathrm{mixture model}^2) \\
     2071  \mathrm{limit}_\mathrm{mixture\ model} &=& 4^2 * (\sigma^2_\mathrm{input} + \sigma_\mathrm{mixture\ model}^2) \\
    20442072  \mathrm{limit}_\mathrm{default} &=& 4^2 * (\sigma^2_\mathrm{input} + (0.1 * \mathrm{value}_\mathrm{input})^2)
    20452073\end{eqnarray}
     
    22692297\end{figure}
    22702298
    2271 
    2272 
    2273 
     2299\section{Difference Images}
     2300\label{sec:diffs}
     2301
     2302Constructing difference images is essentially the same as that used in
     2303the stacking process.  An image is chosen as a template, another image
     2304as the input, and after matching sources to determine the scaling and
     2305transparency, convolution kernels are defined that are used to
     2306convolve one or both of the images to a target PSF.  The images are
     2307then subtracted, and as they should now share a common PSF, static
     2308sources are largely subtracted (completely in an ideal case), whereas
     2309sources that are not static between the two images leave a significant
     2310remnant.  More information on the difference image construction is
     2311contained in \citet{pauls_diff_paper}.  The follow section contains a
     2312overview of the difference image construction used for the data in
     2313DR2.
     2314
     2315The images used to construct difference images can be either
     2316individual warp skycell frames or stacked images, with support for
     2317either to be used as the template or input.  In general, for
     2318differences using stacks, the deepest stack (or the only stack in the
     2319case of a warp-stack difference) is used as the template.  The PV3
     2320processing used warp-stack differences of all input warps against the
     2321stack that was constructed from those inputs.  The same ISIS kernels
     2322as were used in the stack image combination were again used to match
     2323the stack PSF to the input warp PSF.  After convolution of the image
     2324products, the difference is constructed for both the positive (warp
     2325minus stack) and inverse (stack minus warp) to allow for the
     2326photometry of the difference image to detect sources that both rise
     2327and fall relative to the stack.  Note that the convolution process
     2328grows the mask fraction of pixels relative to the warp (the largest
     2329source of masked pixels in these warp stack differences).  Any pixel
     2330that after convolution has any contribution from a masked pixel is
     2331masked as well, ensuring only fully unmasked pixels are used.
     2332
     2333For warp-warp differences, such as those used for the ongoing Solar
     2334System moving object search in nightly observations \citep{MOPS}, the
     2335warp that was taken first is used as the template.  As there is less
     2336certainty in which of the two input images will have better seeing, a
     2337``dual'' convolution method is used.  Both inputs are convolved to a
     2338target PSF that is not identical to either input.  This intermediate
     2339target is essential for the case in which the PSFs of the two inputs
     2340have been distorted in orthogonal directions.  Simply convolving one
     2341to match the other would require some degree of deconvolution along
     2342one axis.  As this convolution method by necessity uses more free
     2343parameters, the ISIS kernels used are chosen to be simpler than those
     2344used in the warp-stack differences.  The ISIS widths are kept the same
     2345(1.5, 3.0, 6.0 pixel FWHMs), but each Gaussian kernel is constrained
     2346to only use a second-order polynomial.  As with the warp-stack
     2347differences, the mask fraction grows between the input warp and the
     2348final difference image due to the convolution.  For the warp-warp
     2349differences, each image mask grows based on the appropriate
     2350convolution kernel, so the final usable image area is highly dependent
     2351on ensuring that the telescope pointings are as close to identical as
     2352possible.  The observing strategy to enable this is discussed in more
     2353detail in \citet{paper1}.
    22742354
    22752355
     
    22972377There is some evidence that we have not fully identified all of these
    22982378crosstalk rules, based on a study of PV3 images.  For example,
    2299 extremely bright stars \czwdraft{exp o5677g0123o has this rule, find a
    2300   magnitude} may be able to create crosstalk ghosts between the second
     2379extremely bright stars %\czwdraft{exp o5677g0123o has this rule, find a
     2380%  magnitude}
     2381may be able to create crosstalk ghosts between the second
    23012382cell column of OTA01 and OTA21, with possibly fainter ghosts appearing
    23022383on OTA11.  Despite the symmetry observed in the main ghost rules,
     
    23292410stacks if fewer pixels need to be rejected.
    23302411
     2412% \czwdraft{one, I believe}
    23312413The fringe model used currently is based on only a limited number of
    2332 days of data \czwdraft{one, I believe}.  This means that the model
    2333 calculated may not be fully sensitive to the exact spectrum of the
    2334 sky.  This may make the model quality differ based on the date and
    2335 local time of observation.  There is some evidence that the fringe
    2336 model does fit some dates better than others, and so improving this by
    2337 expanding the number of input exposures may improve a wider range of
    2338 dates.
     2414days of data.  This means that the model calculated may not be fully
     2415sensitive to the exact spectrum of the sky.  This may make the model
     2416quality differ based on the date and local time of observation.  There
     2417is some evidence that the fringe model does fit some dates better than
     2418others, and so improving this by expanding the number of input
     2419exposures may improve a wider range of dates.
    23392420% o5818g0349o is a good example of bad fringe correction.
    23402421
     
    23492430clip this peak to reduce the noise in the image space is not clear.
    23502431
     2432
    23512433\section{Conclusion}
    23522434
    2353 \czwdraft{Not happy with this.}
     2435%\czwdraft{Not happy with this.}
    23542436
    23552437The Pan-STARRS1 PV3 processing has reduced an unprecidented volume of
    2356 image data, and has produced a catalog of \czwdraft{N} individual
    2357 measurements of \czwdraft{Y} astronomical objects.  Accurately
    2358 calibrating and detrending is essential to ensuring the quality of
    2359 these results.  The detrending process detailed here produces
    2360 consistent data, despite the many individual detectors and their
    2361 individual response functions.
     2438image data, and has produced a catalog for the $3\Pi$ Survey
     2439containing hundreds of billions of individual measurements of
     2440\czwdraft{five billion} astronomical objects.  Accurately calibrating
     2441and detrending is essential to ensuring the quality of these results.
     2442The detrending process detailed here produces consistent data, despite
     2443the many individual detectors and their individual response functions.
    23622444
    23632445From these individual exposures, we are able to construct images on
     
    23672449data set that is ideal for use as a template for image differences.
    23682450
    2369 The Pan-STARRS1 Surveys (PS1) have been
    2370 made possible through contributions by the Institute for Astronomy, the
    2371 University of Hawaii, the Pan-STARRS Project Office, the Max-Planck
    2372 Society and its participating institutes, the Max Planck Institute for
    2373 Astronomy, Heidelberg and the Max Planck Institute for Extraterrestrial
    2374 Physics, Garching, The Johns Hopkins University, Durham University,
    2375 the University of Edinburgh, the Queen's University Belfast, the
    2376 Harvard-Smithsonian Center for Astrophysics, the Las
    2377 Cumbres Observatory Global Telescope Network Incorporated, the
    2378 National Central University of Taiwan, the Space Telescope Science Institute, and the National
    2379 Aeronautics and Space Administration under Grant No. NNX08AR22G issued
    2380 through the Planetary Science Division of the NASA Science Mission
    2381 Directorate, the National Science Foundation Grant No. AST-1238877,
    2382 the University of Maryland, Eotvos Lorand University (ELTE),
    2383 and the Los Alamos National Laboratory.
     2451The Pan-STARRS1 Surveys (PS1) have been made possible through
     2452contributions by the Institute for Astronomy, the University of
     2453Hawaii, the Pan-STARRS Project Office, the Max-Planck Society and its
     2454participating institutes, the Max Planck Institute for Astronomy,
     2455Heidelberg and the Max Planck Institute for Extraterrestrial Physics,
     2456Garching, The Johns Hopkins University, Durham University, the
     2457University of Edinburgh, the Queen's University Belfast, the
     2458Harvard-Smithsonian Center for Astrophysics, the Las Cumbres
     2459Observatory Global Telescope Network Incorporated, the National
     2460Central University of Taiwan, the Space Telescope Science Institute,
     2461and the National Aeronautics and Space Administration under Grant
     2462No. NNX08AR22G issued through the Planetary Science Division of the
     2463NASA Science Mission Directorate, the National Science Foundation
     2464Grant No. AST-1238877, the University of Maryland, Eotvos Lorand
     2465University (ELTE), and the Los Alamos National Laboratory.
    23842466
    23852467
Note: See TracChangeset for help on using the changeset viewer.