Scroll to navigation

ymira(1) MiRA image reconstruction algorithm ymira(1)

NAME

ymira - reconstruct an image from optical interferometric data

SYNOPSIS

ymira [OPTION]... INPUT... OUTPUT

DESCRIPTION

MiRA is a software to reconstruct images from optical interferometric data. The algorithm performs a regularized fit of the data with additional constraints such as normalization and lower and/or upper bounds.

INPUT... are the input OI-FITS data file(s) and OUTPUT is the result saved into a FITS file.

OPTIONS

Options are introduced by a word prefixed with one or two hyphens (at your convenience). In case the first input file starts with an hyphen, a double hyphen can be used to indicate the end of the options.

Some options have units. Angles must be suffixed with mas, milliarcsec, milliarcsecond, milliarcseconds, deg, degree, degrees, rad, radian or radians. Lengths must be suffixed with m, meter, meters, mm, millimeter, millimeters, µm, micron, microns, micrometer, micrometers, nm, nanometer or nanometers.

Some options may take several values which are separated by commas.

Data Selection

The following options can be used to select which data to consider. If no spectral range is given (either with options -effwave and -effband or with options -wavemin and -wavemax), an image is reconstructed from all available wavelengths.
-target=name
Specify the name of the astrophysical object. This option is used to select a specific target from the input file(s). If there is a single target which is identical in all input files, this option may be omitted.
-effwave=length
Specify the effective wavelength; for example, -effwave=1.6micron. This is the central wavelength of the spectral range to consider.
-effband=length
Specify the effective bandwidth; for example, -effband=200nm. This the full width of the spectral range to consider.
-wavemin=length
Specify the minimum wavelength; for example, -wavemin=1.5µm. This the lower bound of the spectral range to consider.
-wavemax=length
Specify the maximum wavelength; for example, -wavemin=1.7µm. This the upper bound of the spectral range to consider.

Image Settings

At least the pixel size (with option -pixelsize) and the image dimensions (with options -fov or -dim) must be specified.
-pixelsize=angle
Specify the angular size of pixels. For example, -pixelsize=0.1mas.
-fov=angle
Specify the angular size of the field of view. For example, -fov=20mas.
-dim=number
Specify the number of pixels per side of the image. For example, -dim=200.
-normalization=value
Specify the sum of pixels. A recommended setting is: -normalization=1.
-min=lower
Specify the lower bound for the pixel values. A recommended setting is:-min=0.
-max=upper
Specify the upper bound for the pixel values.

Direct Model

-xform=name
Specify the method to compute the nonequispaced Fourier transform. Argument name can be nfft, exact or fft (the latter value is however not recommended).

Regularization Settings

-regul=name
Specify the name of the regularization method.
-mu=value(s)
Specify the global regularization weight(s) µ. If a single value is given, it will be used for all reconstructions. If two values are given (separated by comas), a different value of µ will be used for each reconstructions with the values of µ spaced at equal ratios from start to stop (that is, equally spaced logarithmically).
-tau=value
Specify the edge preserving threshold. value must be strictly positive.
-eta=value(s)
Specify the gradient scales along dimensions.
-gamma=angle
Specify the a priori full half width at half maximum (FWHM). For example: -gamma=15mas.

Initial Image

-initial=name
Specify the FITS file or method for initial image.
-seed=value
Specify the seed for the random generator.

Reconstruction Strategy

-bootstrap=count
Specify the number of bootstrapping iterations.
-recenter
Recenter result of bootstrapping iterations.
-threshold=fraction
Specify the level for soft-thresholding input image(s).

Information and Graphics

-quiet
Suppress most messages.
-verb=count
Set the verbose level. Information will be printed every count iteration. Also see option -view.
-view=bits
Bitwise value to specify which graphics to show each time some information is printed. The least significant bits of bits indicate which sub-panel to display. Use -view=0 to suppress all graphics and -view=-1 to show all graphics. The default is to not display anything. If some graphics are shown, the graphic window has to be closed at the end of the reconstruction to effectively quit the program.

Optimizer Settings

-mem=number
Specify the number of previous steps to memorize in VMLMB.
-ftol=real
Specify the function tolerance for the global convergence.
-gtol=real
Specify the gradient tolerance for the global convergence.
-maxiter=number
Specify the maximum number of iterations for all reconstructions.
-maxeval=number
Specify the maximum number of evaluations of the objective function for all reconstructions.

Line Search Parameters

-sftol=real
Specify the function tolerance for the line search.
-sgtol=real
Specify the gradient tolerance for the line search.
-sxtol=real
Specify the step tolerance for the line search.

Output File

-overwrite
Overwrite output file if it exists.
-bitpix=number
Specify the bits per pixel for the output FITS file. Default is -32, that is 32-bit floating point values.
-save_initial
Save initial image as a secondary HDU in the output file.

Miscellaneous

-help
Print out this help.
-version
Print out version number.

AUTHOR

Éric Thiébaut <https://github.com/emmt/MiRA>

REFERENCES

Thiébaut, É.: MiRA: an effective imaging algorithm for optical interferometry, in SPIE Proc. Astronomical Telescopes and Instrumentation 7013, 70131I-1-70131I-12 (2008) <http://dx.doi.org/10.1117/12.788822>

Thiébaut, É. & Giovannelli, J.-F.: Image Reconstruction in Optical Interferometry, in IEEE Signal Processing Magazine 27, pp. 97-109 (2010) <http://dx.doi.org/10.1109/MSP.2009.934870>

23 january 2017 Version 1.1.0