ARCDISP-Fit polynomial dispersion curve.

Usage:

arcdisp in order

Description:

This routine fits a polynomial dispersion curve to a list of identified arc features and transforms the NDF pixel coordinates to spectroscopic values. Optionally you can use a graphical dialogue to improve on the previous feature identification, until you like the appearance of the dispersion curve.

Parameters:

DIALOG

DIALOG = _CHAR (Read) If this is 'Y', 'T' or 'G', then the graphical dialogue is entered before the polynomial dispersion curve for any row is accepted and applied. If this is 'N' or 'F' then the dialogue is not entered and separate dispersion curves are applied to all rows. The string is case-insensitive. ['G']

IN

IN = NDF (Read) The spectrum or set of spectra in which emission features are to be located. This must be a base NDF, the spectroscopic axis must be the first axis. No spectroscopic values or widths must exist in the Specdre Extension. The pixel centres along the first axis must be NDF pixel coordinates. Update access is necessary, the results structure in the Specdre Extension may be modified, an array of spectroscopic values will be created in the Specdre Extensions.

ORDER

ORDER = _INTEGER (Read) The polynomial order of dispersion curves. This cannot be changed during the graphical dialogue. Neither can it differ between rows. [2]

FDB

FDB = NDF (Read) The feature data base. Only the simple list of values FTR_WAVE is used and only in graphics dialogue. It serves to find the identification for an as yet unidentified - but located feature.

DEVICE

DEVICE = GRAPHICS (Read) The graphics device to be used. This is unused if DIALOG is false.

WRANGE

WRANGE( 2 ) = _REAL (Read) In graphical dialogue this parameter is used repeatedly to get a range of laboratory values. This is used for plotting as well as for finding identifications in the feature data base.

Source comments:

   A R C D I S P

   The input data must be a base NDF. They can be a single spectrum
   or a set of spectra. Examples for the latter are a long slit
   spectrum, a set of extracted fibre spectra, or a collapsed
   echellogram (a set of extracted orders from an echelle
   spectrograph). It is necessary that the spectroscopic axis be the
   first axis in the data set. It does not matter how many further
   axes there are, the data will be treated as a linear set of rows
   with each row a spectrum.

   The actual input is the results structure in the Specdre
   Extension. This must be a set of components of type 'arc
   feature'. Each must have two parameters 'centre' and 'laboratory
   value'. These must be corresponding locations one expressed in
   NDF pixel coordinates, the other in spectroscopic values
   (wavelength, frequency etc.). The centres must be strictly
   monotonically increasing, their variances must be available.
   Laboratory values may be bad values to signify unidentified
   features.

   In the graphical dialogue the results structure may be updated.
   The locations remain unchanged; all located features form a fixed
   list of potentially identified features. Identifications may be
   added, deleted or modified. The user has to work on each row in
   turn (unless Quit is chosen). When the user switches from one row
   to the next, the dispersion curve for the finished row is applied
   and its spectroscopic values in the Specdre Extension are set.
   When the last row is finished, the application exits; the output
   of this routine is (i) an updated list of identifications in the
   results structure of the Specdre Extension and (ii) an array of
   spectroscopic values according to the dispersion curves for each
   row, also in the Specdre Extension. At any point the user can
   quit. In this case the array of spectroscopic values is
   discarded, but the updated identifications are retained. If run
   without dialogue, this routine simply performs the polynomial fit
   of the dispersion curve for each row in turn and works out the
   array of spectroscopic values. The list of identifications is
   input only and remains unchanged. If for any row the fit cannot
   be performed, then the spectroscopic values calculated so far are
   discarded and the routine aborts.

   There must not yet be any spectroscopic value information: There
   must be no array of spectroscopic values or widths in the Specdre
   Extension. The pixel centre array for the spectroscopic axis
   (i.e. the first axis) must be NDF pixel coordinates (usually 0.5,
   1.5, ...).

   This routine works on each row (spectrum) in turn. It fits a
   polynomial to the existing identifications. In the optional
   graphical dialogue two plots are displayed and updated as
   necessary. The lower panel is a plot of laboratory values
   (wavelength, frequency etc.) versus pixel coordinate shows

   -  all possible identifications from the feature data base as
      horizontal lines,
   -  all unidentified located features as vertical lines,
   -  all identified located features as diagonal crosses,
   -  the dispersion curve.

   In the upper panel, a linear function is subtracted so that it
   displays the higher-order components of the dispersion curve.
   Crosses indicate the identified located features. Since the scale
   of this upper panel is bigger, it can be used to spot outlying
   feature identifications. In the dialogue you can
      R - Switch to next row, accepting the current fit for this row
      X - X-zoom 2x on cursor
      Y - Y-zoom 2x on cursor
      W - Unzoom to show whole row
      N - Pan by 75% of current plot range
      A - Add ID for location nearest to cursor (from FDB)
      S - Set ID for location nearest to cursor (from cursor y pos.)
      D - Delete ID for feature nearest to cursor
      Q - Quit (preserves updated IDs, discards applied fits for all
          rows)
      ? - Help

   Whenever the list of identifications is changed, the dispersion
   curve is fitted again and re-displayed. If there are too few
   identifications for the order chosen, then the dialogue will
   display the maximum order possible. But such an under-order fit
   cannot be accepted, the R option will result in an error.

   The Q option will always result in an error report, formally the
   routine aborts. After all, it does not achieve the main goal of
   applying individual dispersion curves to all rows.

   On one hand the output of this routine may be an updated list of
   identifications, which could in principle be used in a future run
   of this routine. On the other hand this routine will always
   result in an array of spectroscopic values. The existence of
   these spectroscopic values prevents using this routine again.
   Before using this routine again on the same input NDF you have to
   delete the SPECVALS component in the Specdre Extension.

   In order to facilitate repeated use of this routine on the same
   data, it always uses the Specdre Extension to store spectroscopic
   values, even if the data are one-dimensional and the first axis
   centre array would suffice to hold that information. This leaves
   the first axis centre array at NDF pixel coordinates, as
   necessary for re-use of this routine.

Notes:

This routine recognises the Specdre Extension v. 0.7.

This routine works in situ and modifies the input file.