API Reference

SingleLayer([T, ]size, r0[; interpolate, interpolate_from, size_heuristics=1024])

An AtmosphereSpec that produces independent (uncorrelated) phase frames for each timestep.

Arguments

• T: the number type for phase screens (default Float64).
• size: a tuple (nx, ny) specifying the phase screen shape in pixels (coarse sampler grid).
• r0: Fried parameter (r₀) in pixels.

Keyword Arguments

• interpolate: when specified, the phase screen is sampled at a lower resolution and then upsampled using specified number of Harding interpolation passes. If set to :auto, the number of passes is chosen such that the low-res grid has at most size_heuristics total pixels.
• interpolate_from: alternatively, specify the low-res grid size directly. This must be greater than (11, 11) in each dimension.
• size_heuristics: when interpolate=:auto, the maximum allowed number of pixels in the low-res grid. Tweak this based on the capability of your hardware to compute eigen of a N×N matrix, where N is the number of pixels in the low-res grid.

Notes

The Harding interpolation follows "Fast simulation of a Kolmogorov phase screen" Cressida M. Harding, Rachel A. Johnston, and Richard G. Lane, APPLIED OPTICS Vol. 38, No. 11, April 1999

kolmogorov_covmat(W)
kolmogorov_covmat([T, ]size)

Compute the phase covariance matrix of a turbulent layer in the atmosphere, following the Kolmogorov model. The piston term is excluded in this model. This function assumes unit Fried parameter $r_0 = 1 px$.

Arguments

• T: number type for the covariance matrix (default matches W if provided, otherwise Float64).
• W: the aperture function as a 2D array of weights. Normalized to sum(W) == 1 internally.
• size: a tuple (nx, ny) specifying the size of the aperture function.

DoubleSystem(rel_position, intensity)

Model for a two-component source (binary): primary plus a secondary offset by rel_position. The photon budget and background are configured via the ImagingSpec.photon_count field. See ImagingSpec for details.

Arguments

• rel_position: (dx, dy) integer tuple specifying the secondary's pixel offset.
• intensity: multiplicative intensity of the secondary relative to the primary.

FilterSpec

Representation of a spectral filter used by the imaging pipeline.

FilterSpec(base_wavelength, wavelengths[, intensities])

Arguments

• base_wavelength: central wavelength for the filter.
• wavelengths: vector of sampled wavelengths within the filter bandpass.
• intensities: vector of relative intensities at each sampled wavelength. If not provided, equal weights are assumed.

FilterSpec(base_wavelength; bandwidth[, tcenter=1, tedge=1, npts=7])

Arguments

• base_wavelength: central wavelength for the filter (same units as wavelengths).

Keyword Arguments

• bandwidth: total width of the filter bandpass in wavelength units.
• tcenter: relative intensity at the center wavelength (default 1).
• tedge: relative intensity at the edges of the bandpass (default 1).

ImagingSpec

Container for the imaging system configuration. It is defined by the telescope aperture, the source brightness via photon_count, an optional spectral filter_spec, and the output img_size. If img_size does not match the aperture’s Nyquist grid, the aperture is zero-padded accordingly.

ImagingSpec([T, ]aperture, photon_count[; filter_spec, nyquist_oversample, img_size])
ImagingSpec([T, ]aperture; nphotons, [background, filter_spec, nyquist_oversample, img_size])

Create an imaging system specification.

Arguments

• T: desired number type, inferred from aperture if not provided.
• aperture: 2D aperture (pupil) array describing the telescope pupil.
• photon_count: PhotonCount instance describing the photon budget and background.

Keyword Arguments

• filter_spec: FilterSpec describing sampled wavelengths and their relative intensities. Defaults to a monochromatic filter.
• nyquist_oversample: multiplicative factor applied to the default Nyquist image size (2 * size(aperture)). Defaults to 1. Ignored if img_size is provided.
• img_size: explicit output image size (nx, ny). If not provided, computed from aperture size and nyquist_oversample.
• nphotons and background: alternative way to specify photon budget and background when photon_count is not provided.

PhotonCount(nphotons[, background])

Specifies the photon budget for imaging simulations. Set nphotons to Inf for continuous flux (background can be omitted in this case).

PointSource()

Simple true-sky brightness model. The photon budget and background are configured via the ImagingSpec.photon_count field. See ImagingSpec for details on configuring photon budget and background.

TrueSkyImage(true_sky::AbstractMatrix{T})

Wrap a real-valued true-sky image for use with the imaging pipeline. The photon budget and background are configured via the ImagingSpec.photon_count field. See ImagingSpec for details.

Arguments

• true_sky: real image array representing spatial sky brightness.

CircularAperture([T, ]sz, radius[; aa_dist=1])

Create a circular (optionally anti-aliased) aperture array of shape sz. Returns a 2D numeric array suitable for use as an aperture in ImagingSpec.

Arguments

• T: desired number type, Float64 by default.
• sz: aperture size (nx, ny).
• radius: radius of the circular aperture in pixels. Defaults to the largest that fits.
• aa_dist: anti-aliasing transition width in pixels at the aperture edge.

simulate_images([T, ]img_spec::ImagingSpec, atm_spec::AtmosphereSpec[, true_sky::TrueSky]; n, [batch, filename, verbose, savephases, deviceadapter])

Simulate n images using the provided imaging and atmosphere specifications and write the results to an HDF5 file.

Arguments

• T: output image numeric type; if not provided, defaults to Int for finite-photon simulations (determined by img_spec.photon_count.nphotons) and Float64 for infinite-photon models.
• img_spec: an ImagingSpec describing the aperture, image size, photon budget and filter.
• atm_spec: an AtmosphereSpec used to produce phase screens.
• true_sky: a TrueSky model (e.g. PointSource, DoubleSystem, TrueSkyImage).

Keyword Arguments

• n: number of images to simulate.
• batch: batch size for buffered computations and HDF5 writes (default 128). - filename: output HDF5 filename. If set to nothing (default), no file is written and the images and phases are returned as a NamedTuple of arrays.
• verbose: show progress meter (true by default).
• savephases: when true, the sampled phase screens are saved in the HDF5 in dataset with key "phases", and the pupil function is saved under key "aperture" (true by default).
• deviceadapter: adapter for device-backed arrays (defaults to Array). To use GPU arrays, pass e.g. CUDA.CuArray here (requires CUDA.jl).

simulate_phases(atm_spec::AtmosphereSpec; n, [batch, filename, verbose, deviceadapter])

Simulate n phase screens using the provided atmosphere specification and write the results to an HDF5 file.

Arguments

• atm_spec: an AtmosphereSpec used to produce phase screens.

Keyword Arguments

• n: number of phase screens to simulate.
• batch: batch size for buffered computations and HDF5 writes (default 128).
• filename: output HDF5 filename. If set to nothing (default), no file is written and the phases are returned as an array.
• verbose: show progress meter (true by default).
• deviceadapter: adapter for device-backed arrays (defaults to Array). To use GPU arrays, pass e.g. CUDA.CuArray here (requires CUDA.jl).