Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stecf.org/conferences/sardinia/node13.html
Дата изменения: Mon Jun 12 19:04:27 2006
Дата индексирования: Mon Oct 1 20:55:32 2012
Кодировка:

Поисковые слова: п п п п п п п п п п п п п п
ST-ECF Grism Spectrum Software for NICMOS Grism Data next up previous index
Next: Imaging Polarimetry with NICMOS Up: NICMOS Data Calibration and Previous: The Photometric Performance of

Subsections

ST-ECF Grism Spectrum Software for NICMOS Grism Data

N. Pirzkal, W. Freudling Space Telescope - European Coordinating Facility, Karl-Schwarzschild-Str. 2, D-85748 Garching, Germany

   

 

Abstract:

A unique capability of NICMOS is its grism mode which permits slitless spectrometry at low resolution. Extracting spectra from a large number of NICMOS grism images requires a convenient interactive tool which allows a user to manipulate direct/grism image pairs. NICMOSlook and Calnic-C are IDL programs designed for that purpose at the Space Telescope - European Coordinating Facility. NICMOSlook is a graphical interface driven version of this extraction tool, while Calnic C is a program which performs the same functionality in a ``pipeline`` approach.

NICMOS, grism, spectral extraction

NICMOSlook and Calnic C

NICMOSlook and Calnic-C are two programs that were designed to take advantage of the fact that the NICMOS grism mode typically consists of taking both a direct and a grism image. Both NICMOSlook and Calnic-C use the direct image to determine which objects are to be extracted, their location, size, and orientation. The programs allow the user to extract spectra using a range of extraction techniques.
NICMOSlook, which has an IDL user interface based on STISlook from Terry Beck, is designed to allow a user to interactively and efficiently extract spectra from a small batch of data. The user maintains full control over the extraction process. NICMOSlook was designed to be most useful to users who want to quickly examine data or fine tune the extraction process.
The stand-alone version, Calnic C, is meant to be used on larger sets of data and requires only a minimum amount of user input. Both programs can be fully configured and extended through the use of a few configuration and calibration files. Figure 1 presents an overview of the spectral extraction process.

Features overview

Image Display

NICMOSlook provides a variety of display options to examine the direct and grism images (Figure 2). These options include color tables, zoom factors, row/cut plots, and blinking between the two images. Some basic image filtering and processing capabilities are also available to facilitate the identification of objects.
Calnic-C outputs Postscript finding charts which contain the direct and grism images with the objects and spectra clearly numbered and marked (Figures 4 and 5).

Object Identification

With NICMOSlook, objects can be identified automatically using either the DAOFIND algorithm, by using a text file containing the objects coordinates, or can be provided interactively by the user by marking specific objects in the direct image. NICMOSlook and Calnic-C account for the size and orientation of the objects when extracting their spectrum (Figure 3).
Calnic-C uses the more sophisticated SExtractor (Bertin 1996) program to automatically locate and determine the size and orientation of objects in the direct images. Calnic-C can also be made to extract only objects that are listed in an input list.
With both programs, and if using an input object list, the RA and DEC of objects can be used rather than simply the pixel coordinates of the objects in the direct image. This feature requires that a WCS be available in both the direct and grism images FITS headers.

Determination of the location of the spectrum

The positions of the objects are used in combination with the appropriate coordinate transformation to determine the location and extend of spectra in the grism image. The zeroth, first, and second spectral orders can be automatically located on the grism image (Figure 3). The direct and grism images do not need to have been taken with exactly the same telescope pointing since NICMOSlook and Calnic-C can additionally make use of the WCS information to locate the spectra in the grism image.

Spectral Extraction

Weighted or unweighted spectral extractions can be performed. Objects can be extracted as point sources (In which case, a Tiny Tim profile is used during the spectral extraction) or as extended objects (In which case a spectral profile is determined by first computing the profile of the object in the direct image).

Background Subtraction

The programs estimate the amount of background light in a spectrum. A region containing no spectrum is first identified in the vicinity of the spectrum of interest (Figure 6). This region is then either used to scale a pre-computed model background, or is linearly interpolated to estimate the level of the background were the spectrum is located.

Wavelength and Flux calibrations

A wavelength is assigned to each pixel of the spectrum in the grism image using the knowledge of where the object is located in the direct image and using the dispersion relation of the grism used. A flatfield is then constructed by taking into account that different pixels in the grism image correspond to different wavelengths. NICMOSlook and Calnic-C compute the flatfield coefficient of a specific pixel at a specific wavelength by interpolating the narrow band filter flatfield coefficients of the same pixel. Inverse sensitivity curves are then applied to the extracted spectra to produce a final extraction product that is wavelength and flux calibrated (Figures 7 and 8).

Deblending

An attempt can be made to automatically remove the spectral contamination caused by the presence of nearby objects. The only assumption made by the method used is that the shape of the object is not wavelength dependent.

Spectral lines search

Emission and absorption lines are automatically identified in the resulting spectra, and the continuum emission is automatically determined.

Result Output

The final data products are plots on the screen (Figures 7 and 8), binary FITS tables and postscript files with the spectra, error estimates, object parameters derived from the direct imaging, and details of the spectrum extraction process (Figure 9). Calnic-C additionally outputs finding charts to locate the objects and their spectra in the direct and grism images

  
Figure 1: Overview of NICMOSlook and Calnic-C
\begin{figure}\psfig{file=pirzkaln1.eps,width=4.0in}
\end{figure}





  
Figure 2: The NICMOSlook user interface
\begin{figure}\psfig{file=pirzkaln2.ps,width=4.5in}
\end{figure}





  
Figure 3: Grism image displayed by NICMOSlook with the identified objects marked. First order spectra are marked with full lines, second order spectra are marked with dashed lines
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln3.ps,width=3.0in}
\end{figure}


  
Figure 4: Direct image Postscript finding chart output by Calnic-c
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln4.ps,width=3.0in}
\end{figure}


  
Figure 5: Grism image Postscript finding chart output by Calnic-c
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln5.ps,width=3.0in}
\end{figure}


  
Figure 6: Display of the areas used to determine the background spectral contamination
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln6.ps,width=3.0in}
\end{figure}


  
Figure 7: Result of a first order extraction of Vy-22
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln7.ps,width=3.0in}
\end{figure}


  
Figure 8: Result of a second order extraction of Vy-22
\begin{figure}\hspace{1.0in}
\psfig{file=pirzkaln8.ps,width=3.0in}
\end{figure}


  
Figure 9: Postscript output of Calnic-C
\begin{figure}\psfig{file=pirzkaln9.ps,width=4.5in,angle=-90}
\end{figure}

Supported platforms and availability


\begin{references}% latex2html id marker 2097
\reference Bertin, E., Arnouts, S., 1996, \aap, 117, 393
\end{references}


next up previous index
Next: Imaging Polarimetry with NICMOS Up: NICMOS Data Calibration and Previous: The Photometric Performance of
Norbert Pirzkal
1998-07-09