MOSAIC
Merging intersecting fields

• Coordinates recentering
• Truncate datacubes
• Merge datacubes

Coordinates recentering

• Principles
  
  To be able to mosaic fields, that is to be able to merge datacubes obtained from different exposures, the exposures must, first, belong to the same scenario. Second, one must know exactly the spatial shift between the exposures, to avoid any field distorsion in the resulting datacube. This shift has to be evaluated by the user. The usual way is the following :
  
  • Use the Analyse / Integrate Spectra function to produce, for each exposure to be merged, an image of the field integrated over the whole spectral range.
  • On these images, find some eye-catching detail (a star is best, but any peaked maximum will do the job, or at least some small morphological feature), and note the [alpha_center,delta_center] position of this point. One may use the Display / Datacube explorer, and note visually the displayed coordinates, or use the Image / Fit PSF function. You may want to use then the Display result function to have a look at the image with the fitted point overlaid.
    The point thus defined must be present on ALL the images, that is it must belong to the intersection of the reconstructed images. If some exposure have an empty intersection with some other ones, it may not be possible to merge it, except if it can be first properly merged with another exposure which in turn intersects the rest of the data.
  • On each exposure to be later merged, this particular point will become the [alpha=0,delta=0] point, after the completion of the coordinates recentering function.


• Use
  
  
  • In the main menu, click on Mosaic, then on Coordinates recentering.
  • In the new window which pops up, enter the name of the Tiger datacube to be recentered, and the point which should now be considered as the Center of the field.
  • Click on Accept.


• Standard parameters
  
  
  • Tiger datacube : the name of the datacube to be recentered; it will keep its name after the operation (which in fact affects only the associated table). May be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in.
  • Center : in these fields, enter the [alpha_center,delta_center] you just obtained on the images reconstructed from this datacube (see Principles above) for the common reference point.


• Optional parameters
  
  
  • Debug : This switches the program to verbose mode, and more informations are recorded into the history file (see Getting started).
    
  • Save values : All the input values (datacube name, coordinates, ...) are saved, and become the new default values for this user. They can be recalled at will, and are used each time the Coordinates recentering window is opened.
    
  • Recall values : The values (datacube name, coordinates, ...) saved by the user are loaded to the various input fields.
    
  • Default values : The input fields are set to the general defaults values; for instance, the datacube name is set to blank.
    

Truncate datacubes

• Principle
  
  The datacubes to be merged must hold spectra which share the same wavelength range. If you went all the way through the complete reduction process, the flat (and/or the flux) calibration has done the job, and all the spectra are already truncated to the same wavelength limits. But if (and only if) you want to merge datacubes which have not been flat (or flux) -calibrated, then you must use the present function.
  Spectra which do not contain entirely the wavelength range are discarded in the new datacube.



• Use
  
  
  • In the main menu, click on Mosaic, then on Truncate datacube.
  • In the new window which pops up, enter the names of the Input datacube and Output datacube.
  • Choose the wavelength range to be kept, either as a couple of numerical values, or to be taken from some template spectrum.
  • Click on Accept.


• Standard parameters
  
  
  • Input datacube : the name of the datacube to be lambda-truncated. May be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in.
  • Output datacube : the name of the resulting truncated datacube. May be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in.


• Optional parameters
  
  
  • Wavelength range : give here the wavelength limits (Å) for the truncate operation. Keep in mind that "short" spectra, which do not contain entirely the wavelength interval, will be discarded.
  • Get wavelength range from spectrum : you may decide to take some particular spectrum as a template for the operation. The name of the spectrum may be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in.
  • Display removed lenses only : by checking this button, you get a display of the discarded lenses.
  
  
  • Save values : All the input values (datacube names, wavelengths, ...) are saved, and become the new default values for this user. They can be recalled at will, and are used each time the Truncate datacubes window is opened.
    
  • Recall values : The values (datacube names, wavelengths, ...) saved by the user are loaded to the various input fields.
    
  • Default values : The input fields are set to the general defaults values; for instance, the datacube names are set to blank.
    

Merge datacubes

• Principles
  
  A new datacube is created, holding the reunion of the data contained in the datacubes merged (under some conditions, see below). The new datacube has a square spatial sampling grid, instead of the original hexagonal grid.
  


• Use
  
  
  • In the main menu, click on Mosaic, then on Merge datacubes.
  • In the new window which pops up, enter the names of the Input datacubes. There is no a priori limit on the number of input datacubes.
  • Give the name of the resultant Output datacube.
  • Click on Accept.


• Standard parameters
  
  
  • Input datacubes : the names of the datacubes to be merged. They may be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in. You may specify a Weight factor, if you think the data quality is widely different among exposures. You may also specify a Normalization factor, for the case you are about to merge exposures which are not flux-calibrated, and have, for instance, different exposure times. Usually, accept the unit default value for both.
  • Output datacube : the name for the datacube resulting from the merge action. It may be dragged and dropped from the Reduction folder, selected using the browse icon at the end of the field, or just typed in.


• Optional parameters
  
  
  • Output spatial sampling : it may be set to Auto, and it is computed after the spatial sampling of the original datacubes, or set to the desired value (in arcsecond).
  • Output spatial limits : If unchecked, the whole spatial region where valid spectra can be defined is used. If not, you have to specify the [X1,Y1] and [X2,Y2] bottom left and top right angles of the rectangular region you want to cover.
  • Merge options :
    • Adjust center : if you check that, a spectrum will be created by interpolation at the [0,0] point, and the grid will be built around this.
    • Bicubic interpolation : once checked, the interpolation process is no longer bilinear, but bicubic. Better, specially for high-gradient regions, but time-consuming, and needs a good S/N ratio; otherwise, the algorithm may be fooled by local noisy gradients.
  • Merge method :
    • Average : the weighted average is used for the interpolation of spectral data.
    • Median : the median is used for interpolation; this needs a lot of high S/N ratio exposures to work.
  • Minimum number of spectra : you may ask to discard a point if there are not at least N neighbouring spectra to interpolate; but usually, people want to keep every point where there is at least one spectrum nearby... So, 1 is a good default value.
  • Debug
    This switches the program to verbose mode, and more informations are recorded into the history file (see Getting started).
    
  • Save values : All the input values (datacube names, ...) are saved, and become the new default values for this user. They can be recalled at will, and are used each time the Merge datacubes window is opened.
    
  • Recall values : The values (datacube names, ...) saved by the user are loaded into the various input fields.
    
  • Default values : The input fields are set to the general defaults values; for instance, the datacube names are set to blank.
    

Back to top of chapter.

Back to OASIS home page.