REGRID - Applies a geometrical transformation to an NDF

Description:

This application uses a specified Mapping to re-grid the pixel positions in an NDF . The specified Mapping should transform pixel co-ordinates in the input NDF into the corresponding pixel co-ordinates in the output NDF.

By default, the bounds of the output pixel grid are chosen so that they just encompass all the transformed input data, but they can be set explicitly using Parameters LBOUND and UBOUND.

Two algorithms are available for determining the output pixel values: resampling and rebinning (the method used is determined by the REBIN parameter).

The Mapping to use can be supplied in several different ways (see Parameter MAPPING).

Usage:

regrid in out [method]

Parameters:

CONSERVE = _LOGICAL (Read)

Only accessed when using the resampling algorithm (i.e. if REBIN is set FALSE). If set TRUE, then the output pixel values will be scaled in such a way as to preserve (approximately) the total data value in a feature on the sky. The scaling factor is the ratio of the output pixel size to the input pixel size. This option can only be used if the Mapping is successfully approximated by one or more linear transformations. Thus an error will be reported if it used when the TOL parameter is set to zero (which stops the use of linear approximations), or if the Mapping is too non-linear to be approximated by a piece-wise linear transformation. The ratio of output to input pixel size is evaluated once for each panel of the piece-wise linear approximation to the Mapping, and is assumed to be constant for all output pixels in the panel. Flux conservation can only be approximate when using the resampling algorithm. For accurate flux conservation set REBIN to TRUE. [FALSE]

IN = NDF (Read)

The NDF to be transformed.

LBOUND( ) = _INTEGER (Read)

The lower pixel-index bounds of the output NDF. The number of values must be equal to the number of dimensions in the output NDF. If a null value is supplied, default bounds will be used which are just low enough to fit in all the transformed pixels of the input NDF. [!]

MAPPING = FILENAME (Read)

The name of a file containing the Mapping to be used, or null (!) if the input NDF is to be mapped into its own current Frame. If a file is supplied, the forward direction of the Mapping should transform pixel co-ordinates in the input NDF into the corresponding pixel co-ordinates in the output NDF. The file may be one of the following.

• A text file containing a textual representation of the AST Mapping to use. Such files can be created by WCSADD.

• A text file containing a textual representation of an AST FrameSet. If the FrameSet contains a Frame with Domain PIXEL, then the Mapping used is the Mapping from the PIXEL Frame to the current Frame. If there is no PIXEL Frame in the FrameSet, then the Mapping used is the Mapping from the base Frame to the Current Frame.

• A FITS  file. The Mapping used is the Mapping from the FITS pixel co-ordinates in which the centre of the bottom-left pixel is at co-ordinates (1,1), to the co-ordinate system represented by the primary WCS headers, CRVAL, CRPIX, etc.

• An NDF. The Mapping used is the Mapping from the PIXEL Frame to the Current Frame of its WCS FrameSet.

If a null (!) value is supplied, the Mapping used is the Mapping from pixel co-ordinates in the input NDF to the current Frame in the input NDF. The output NDF will then have pixel co-ordinates which match the co-ordinates of the current Frame of the input NDF (apart from possible additional scalings as specified by the SCALE parameter).

METHOD = LITERAL (Read)

The method to use when sampling the input pixel values (if resampling), or dividing an input pixel value between a group of neighbouring output pixels (if rebinning). For details of these schemes, see the descriptions of routines AST_RESAMPLEx and AST_REBINx in SUN/210. METHOD can take the following values.

• "Bilinear" -- When resampling, the output pixel values are calculated by bi-linear interpolation among the four nearest pixels values in the input NDF. When rebinning, the input pixel value is divided up bi-linearly between the four nearest output pixels. Produces smoother output NDFs than the nearest-neighbour scheme, but is marginally slower.

• "Nearest" -- When resampling, the output pixel values are assigned the value of the single nearest input pixel. When rebinning, the input pixel value is assigned completely to the single nearest output pixel.

• "Sinc" -- Uses the ${\rm sinc}({\pi}x)$ kernel, where $x$ is the pixel offset from the interpolation point (resampling) or transformed input pixel centre (rebinning), and ${\rm sinc}(z)=\sin(z)/z$. Use of this scheme is not recommended.

• "SincSinc" -- Uses the ${\rm sinc}({\pi}x){\rm sinc}(k{\pi}x)$ kernel. A valuable general-purpose scheme, intermediate in its visual effect on NDFs between the bi-linear and nearest-neighbour schemes.

• "SincCos" -- Uses the ${\rm sinc}({\pi}x)\cos(k{\pi}x)$ kernel. Gives similar results to the "Sincsinc" scheme.

• "SincGauss" -- Uses the ${\rm sinc}({\pi}x)e^{-kx^2}$ kernel. Good results can be obtained by matching the FWHM of the envelope function to the point-spread function of the input data (see Parameter PARAMS).

• "Somb" -- Uses the ${\rm somb}({\pi}x)$ kernel, where $x$ is the pixel offset from the interpolation point (resampling) or transformed input pixel centre (rebinning), and ${\rm somb}(z)=2*J_{1}(z)/z$ ($J_1$ is the first-order Bessel function of the first kind. This scheme is similar to the "Sinc" scheme.

• "SombCos" -- Uses the ${\rm somb}({\pi}x)\cos(k{\pi}x)$ kernel. This scheme is similar to the "SincCos" scheme.

• "Gauss" -- Uses the $e^{-kx^2}$ kernel. The FWHM of the Gaussian is given by Parameter PARAMS(2), and the point at which to truncate the Gaussian to zero is given by Parameter PARAMS(1).

• "BlockAve" -- Block averaging over all pixels in the surrounding $N$-dimensional cube. This option is only available when resampling (i.e. if REBIN is set to FALSE).

All methods propagate variances from input to output, but the variance estimates produced by these schemes other than nearest neighbour need to be treated with care since the spatial smoothing produced by these methods introduces correlations in the variance estimates. Also, the degree of smoothing produced varies across the NDF. This is because a sample taken at a pixel centre will have no contributions from the neighbouring pixels, whereas a sample taken at the corner of a pixel will have equal contributions from all four neighbouring pixels, resulting in greater smoothing and lower noise. This effect can produce complex Moiré patterns in the output variance estimates, resulting from the interference of the spatial frequencies in the sample positions and in the pixel-centre positions. For these reasons, if you want to use the output variances, you are generally safer using nearest-neighbour interpolation. The initial default is "Nearest". [current value]

OUT = NDF (Write)

The transformed NDF.

PARAMS( 2 ) = _DOUBLE (Read)

An optional array which consists of additional parameters required by the Sinc, SincSinc, SincCos, SincGauss, Somb, SombCos, and Gauss methods.

PARAMS(1) is required by all the above schemes. It is used to specify how many pixels are to contribute to the interpolated result on either side of the interpolation or binning point in each dimension. Typically, a value of 2 is appropriate and the minimum allowed value is 1 (i.e. one pixel on each side). A value of zero or fewer indicates that a suitable number of pixels should be calculated automatically. [0]

PARAMS(2) is required only by the Gauss, SincSinc, SincCos, and SincGauss schemes. For the SombCos, SincSinc, and SincCos schemes, it specifies the number of pixels at which the envelope of the function goes to zero. The minimum value is 1.0, and the run-time default value is 2.0. For the Gauss and SincGauss scheme, it specifies the full-width at half-maximum (FWHM) of the Gaussian envelope. The minimum value is 0.1, and the run-time default is 1.0. On astronomical images and spectra, good results are often obtained by approximately matching the FWHM of the envelope function, given by PARAMS(2), to the point-spread function of the input data. []

REBIN = _LOGICAL (Read)

Determines the algorithm used to calculate the output pixel values. If a TRUE value is given, a rebinning algorithm is used. Otherwise, a resampling algorithm is used. See the ``Choice of Algorithm'' topic below. [current value]

SCALE( ) = _DOUBLE (Read)

Axis scaling factors which are used to modify the supplied Mapping. If the number of supplied values is fewer than the number of output axes associated with the Mapping, the final supplied value is duplicated for the missing axes. In effect, transformed input co-ordinate axis values would be multiplied by these factors to obtain the corresponding output pixel co-ordinates. If a null (!) value is supplied for SCALE, then default values are used which depends on the value of Parameter MAPPING. If a null value is supplied for MAPPING then the default scaling factors are chosen so that pixels retain their original size (very roughly) after transformation. If as non-null value is supplied for MAPPING then the default scaling factor used is 1.0 for each axis (i.e. no scaling). [!]

TITLE = LITERAL (Read)

A Title for the output NDF structure. A null value (!) propagates the title from the input NDF to the output NDF. [!]

TOL = _DOUBLE (Read)

The maximum tolerable geometrical distortion which may be introduced as a result of approximating non-linear Mappings by a set of piece-wise linear transforms. The resampling algorithm approximates non-linear co-ordinate transformations in order to improve performance, and this parameter controls how inaccurate the resulting approximation is allowed to be, as a displacement in pixels of the input NDF. A value of zero will ensure that no such approximation is done, at the expense of increasing execution time. [0.2]

UBOUND( ) = _INTEGER (Read)

The upper pixel-index bounds of the output NDF. The number of values must be equal to the number of dimensions of the output NDF. If a null value is supplied, default bounds will be used which are just high enough to fit in all the transformed pixels of the input NDF. [!]

WLIM = _REAL (Read)

This parameter is only used if REBIN is set TRUE. It specifies the minimum number of good pixels which must contribute to an output pixel for the output pixel to be valid. Note, fractional values are allowed. A null (!) value causes a very small positive value to be used resulting in output pixels being set bad only if they receive no significant contribution from any input pixel. [!]

Examples:

regrid sg28948 sg28948r mapping=rotate.ast

Here sg28948 is resampled into a new co-ordinate system using the AST Mapping stored in a text file called rotate.ast (which may have been created using WCSADD for instance).

regrid flat distorted mapping=!

This transforms the NDF called flat into its current co-ordinate Frame, writing the result to an NDF called distorted. It uses nearest-neighbour resampling. If the units of the PIXEL and current co-ordinate Frames of flat are of similar size, then the pixel co-ordinates of distorted will be the same as the current co-ordinates of flat, but if there is a large scale discrepancy a scaling factor will be applied to give the output NDF a similar size to the input one. The output NDF will be just large enough to hold the transformed copies of all the pixels from NDF flat.

regrid flat distorted mapping=! scale=1 method=sinccos params=[0,3]

As the previous example, but the additional scaling factor will not be applied even in the case of large size discrepancy, and a sinc$*$cos one-dimensional resampling kernel is used which rolls off at a distance of 3 pixels from the central one.

regrid flat distorted mapping=! scale=0.2 method=blockave params=2

In this case, an additional shrinking factor of 0.2 is being applied to the output NDF (i.e. performed following the Mapping from pixel to current co-ordinates), and the resampling is being done using a block averaging scheme in which a cube extending two pixels either side of the central pixel is averaged over to produce the output value. If the PIXEL-domain and current Frame pixels have (about) the same size, this will result in every pixel from the input NDF adding a contribution to one pixel of the output NDF.

regrid a119 a119s mapping=! lbound=[1,-20] ubound=[256,172]

This transforms the NDF called a119 into an NDF called a119s. It uses nearest-neighbour resampling. The shape of a119s is forced to be (1:256,Â$-$20:172) regardless of the location of the transformed pixels of a119.

Notes:

• If the input NDF contains a VARIANCE component, a VARIANCE component will be written to the output NDF. It will be calculated on the assumption that errors on the input data values are statistically independent and that their variance estimates may simply be summed (with appropriate weighting factors) when several input pixels contribute to an output data value. If this assumption is not valid, then the output error estimates may be biased. In addition, note that the statistical errors on neighbouring output data values (as well as the estimates of those errors) may often be correlated, even if the above assumption about the input data is correct, because of the sub-pixel interpolation schemes employed.

• This task is based on the AST_RESAMPLEx and AST_REBINx routines described in SUN/210.

Choice of Algorithm

The algorithm used to produce the output image is determined by the REBIN parameter, and is based either on resampling the output image or rebinning the input image.

The resampling algorithm steps through every pixel in the output image, sampling the input image at the corresponding position and storing the sampled input value in the output pixel. The method used for sampling the input image is determined by the METHOD parameter. The rebinning algorithm steps through every pixel in the input image, dividing the input pixel value between a group of neighbouring output pixels, incrementing these output pixel values by their allocated share of the input pixel value. The way in which the input sample is divided up between the output pixels is determined by the METHOD parameter.

The two algorithms behaviour quite differently if the transformation from input to output includes any significant change of scale. In general, resampling will not alter the pixel values associated with a source, even if the pixel size changes. On the other hand, the rebinning algorithm will change the pixel values in order to correct for a change in pixel size. Thus, rebinning conserves the total data value within a given region where as resampling, in general, does not (but see the discussion of the CONSERVE parameter below).

Resampling is appropriate if the input image represents the spatial density of some physical value (e.g. surface brightness) because the output image will have the same normalisation as the input image. However, rebinning is probably more appropriate if the image measures (for instance) flux per pixel, since rebinning takes account of the change in pixel size.

Another difference is that resampling guarantees to fill the output image with good pixel values (assuming the input image is filled with good input pixel values), whereas holes can be left by the rebinning algorithm if the output image has smaller pixels than the input image. Such holes occur at output pixels which receive no contributions from any input pixels, and will be filled with the value zero in the output image. If this problem occurs the solution is probably to change the width of the pixel spreading function by assigning a larger value to PARAMS(1) and/or PARAMS(2) (depending on the specific METHOD value being used).

Both algorithms have the capability to introduce artificial artefacts into the output image. These have various causes described below.

• Particularly sharp features in the input can cause rings around the corresponding features in the output image. This can be minimised by suitable settings for the METHOD and PARAMS parameters. In general such rings can be minimised by using a wider interpolation kernel (if resampling) or spreading function (if rebinning), at the cost of degraded resolution.

• Regular patterns of curvy lines covering the whole output image can be created when using the rebinning algorithm. These are caused by a beating effect between the input pixel positions and the output pixel's position, and their nature and strength depend critically upon the nature of the Mapping and the spreading function being used. In general, the nearest-neighbour spreading function demonstrates this effect more clearly than the other functions, and for this reason should be used with caution. Again, wider spreading functions reduce the effect at the cost of degraded resolution. Note, the resampling algorithm is not subject to these artefacts. For this reason, you may prefer to use the resampling algorithm with the CONSERVE parameter set to TRUE. This causes the resampling algorithm to perform approximate flux conservation.

• The approximation of the Mapping using a piece-wise linear transformation (controlled by Parameter TOL) can produce artefacts at the joints between the panels of the approximation. These can occur when using the rebinning algorithm, or when using the resampling algorithm with CONSERVE set to TRUE. They are caused by the discontinuities between the adjacent panels of the approximation, and can be minimised by reducing the value assigned to the TOL parameter.

Related Applications

KAPPA: FLIP, ROTATE, SLIDE, WCSADD, WCSALIGN; CCDPACK: TRANLIST, TRANNDF, WCSEDIT.

Implementation Status:

• The LABEL, UNITS, and HISTORY components, and all extensions are propagated. TITLE is controlled by the TITLE parameter. DATA, VARIANCE, and WCS are propagated after appropriate modification. The QUALITY component is also propagated if Nearest-Neighbour interpolation is being used. The AXIS component is not propagated.

• Processing of bad pixels and automatic quality masking are supported.

• All non-complex numeric data types can be handled. If REBIN is TRUE, the data type will be converted to one of _INTEGER, _DOUBLE or _REAL for processing.

• There can be an arbitrary number of NDF dimensions.