Next: Karma: a Visualization Test-Bed
Previous: Systems Aspects of COBE Science Data Compression
Table of Contents --- Search --- PS reprint


Astronomical Data Analysis Software and Systems V
ASP Conference Series, Vol. 101, 1996
George H. Jacoby and Jeannette Barnes, eds.

MATADOR: Software for the Manipulation of 3D Data

V. Gavryusev

Osservatorio Astrofisico di Arcetri, Firenze, Italy

C. Muñoz-Tuñón

Instituto de Astrofísica de Canarias, La Laguna, Spain

Abstract:

The evaluation of 3D data-cubes obtained by bidimensional spectroscopy is not a very simple task due to the huge amount of data that they contain. A typical data-cube could be 256x256 pixels in the spatial plane and 100 frames in the spectral direction. We present a powerful software package, MATADOR, which can easily handle this kind of data.

MATADOR is written on IDL1 but it is not necessary for the user to have any knowledge of this language. MATADOR offers to the user a widget based interface, and any task is performed by simple manipulation of the mouse. The program is basically interactive, although some time/memory consuming operations can be executed as independent sub-tasks. The available tools can visualize an image at any desired wavelength interval and analyze the spectrum for any selected region on the image. The spectrum can, on the spot, be fitted by the combination of the background and several Gaussian or Lorentzians. The asymmetry parameters of emission-lines can also be studied. Different possibilities exist for manipulating two-dimensional images, such as masking, contouring and combining options. There are also tools for investigating the velocity distribution and the structure function of any chosen area. On-line help and a flexible hard-copy option are available, too.

1. Introduction

The main problem that one has with huge 3D data cubes, such as data obtained from the TAURUS II Fabry-Perot imaging spectrograph (Atherton & Taylor 1980), comes from their large data size. They provide an amount of data that often makes you want to run away before any useful information is in sight. The available software was not powerful enough to match the potential of the instrument, and we decided to design new software that could handle the data. The outcome was named MATADOR (Gavryusev 1994; Muñoz-Tuñón, Gavryusev, & Castaneda 1995).

Although MATADOR was created for manipulating TAURUS spectral line data cubes, it does not contain any instrument-specific features and can be used in the analysis of other similar 3D data.

The program is designed to be interactive, but some time/memory consuming operations are performed as subtasks. While MATADOR has been written in IDL and needs the IDL environment to be installed, it is not necessary for the user to have any knowledge of this language. MATADOR can be executed only in the X Windows environment. Usually it is used as a stand-alone program, but it can be started during the IDL session as any other standard IDL routine.

MATADOR offers the user a widget based interface. Any task is performed by simple and quick manipulation of the mouse; by clicking/pressing/releasing its buttons and/or by dragging it. Only occasionally is it necessary to edit some simple text fields on the widgets.

Below we describe briefly the functionality provided by MATADOR.

2. Functionality

After startup, the program offers to the user 3 basic options within the Main Menu. They are the Set-up of the session, the Read option for reading the 3D data cube into memory, and Tools which provides the access to the list of available facilities for the data analysis.

  
Figure 1: Example of several MATADOR widgets.
Figure 1: PS 1.7 Mb

The basic tool, Collapse (Figure 1), makes it possible to visualize the image of the x,y plane as a sum of all z planes for any desired wavelength (z) interval and to analyze the spectrum corresponding to any selected region on the image. After the choice of the region, the widget with the corresponding averaged or integrated spectrum is created. This spectrum can be immediately fitted by the combination of background and several Gaussian or Lorentzians (from one to any number of them, predefined in the session environment). All fitting parameters are shown in the picture. The x-scale of the spectrum can be changed in the chain: array index (etalon step)---wavelength---velocity---by the simple click of the mouse button on the corresponding widget button. The information on the picture, dependent on the choice of the scale, is changed automatically. For single line fitted spectra it is possible to work out the asymmetry indices, as described by Heckman et al. (1981). There are also tools for the investigation of velocity distribution and the structure function of any chosen area.

Any rectangular part of the x,y plane, including the whole image itself, can be magnified, and this zoomed image is evaluated exactly in the same manner as the original one. Hence, there is no problem manipulating a single pixel or a complicated region constructed from a small number of pixels.

MATADOR can work with several 2D images simultaneously, for example, a collapsed image and corresponding line-position and maps.

If the user likes, he/she can smooth the 3D cube by the convolution with a running centered window, where the smoothing function is pre-set. The window can be 1, 2 or 3 dimensional boxes of arbitrary sizes. At this time the software provides three smoothing functions: simple average, Gaussian and Sinc. It is also possible to perform a Global fitting of the spectra, contained in the 3D data cube, for each point x,y. In this case the single Gaussian in combination with the background is used. It is possible to avoid the fitting for some pixels, but generally it is not necessary because of the presence in MATADOR of such tools as Mask and Combine Images (see below). The output of the global fitting is four 2D maps of the line amplitude, line position, and background of the fitted gaussian, as well as four more maps with associated errors. Any of these maps can be manipulated as a standard collapsed image with complete access to all its functionality. The maps of line position and (and their errors) can be plotted in any desired scale---array index, wavelength or velocity.

There are several other useful tools designed to manipulate two 2D images. The Mask allows one to assign any chosen value to a given interval of image values. That is clearly a way to mask the image or hide a chosen region on the x,y plane. Typical usage of this tool could be, for instance, to set to unity all pixels with line errors below a given limit and to set everything else to zero. Then, with the help of Combine Images, it is possible to build up the set of maps following this condition from the maps obtained by the global fitting without any condition. It takes significantly less time then to make global fittings for each type of conditions. With the Combine Images tool one can perform all algebraic and logical operations between two 2D images, as well as single argument operations like power, sin, cos, exp, ln and log. The number of possible applications for which the combining of Mask and Combine Images could be useful remains open to the user's imagination.

With the Contour tool it is possible to plot the chosen contours (lines with constant value) of a given image into another one (maybe the same) and also to display the image part contained in between two given isocontours.

It is also possible to simulate new 3D data cubes based on the predefined maps of fitting parameters (completely artificial or created from previous global fittings).

Finally, any picture of interest to the user can be printed immediately or stored as a PostScript file for subsequent use (the encapsulated format is supported, too). The place and size of the picture on the paper can be easily changed by the user since everything is predefined to fit the picture on the A4 paper. PostScript and Color PostScript printers are currently supported.

Acknowledgments:

VG acknowledges the hospitality of the IAC that, via the GEFE project, provided all the resources needed for the development of MATADOR. We are grateful to R. Kroll, the IAC Computer Center manager, for his kindness and help in matters related to MATADOR.

References:

Atherton, P. D., & Taylor, K. 1980, MNRAS, 191, 675

Gavryusev, V. 1994, Instituto de Astrofísica de Canarias, MATADOR user's guide

Muñoz-Tuñón, C., Gavryusev, V., & Castaneda, H. O. 1995, AJ, 110, 1630

Heckman, T. M., et al. 1981, ApJ, 247, 403

1IDL is a registered trademark of Research Systems, Inc.


Next: Karma: a Visualization Test-Bed
Previous: Systems Aspects of COBE Science Data Compression
Table of Contents --- Search --- PS reprint
Wed Jul 3 07:44:22 MST 1996