KPG1_GAUSx

Smooths a two-dimensional array using a symmetrical Gaussian PSF

Description:

The routine smooths the array A using a symmetrical Gaussian point-spread function and returns the result in the array B.

Invocation

CALL KPG1_GAUSx( SIGMA, IBOX, SAMBAD, WLIM, NX, NY, BAD, VAR, A, B, BADOUT, WEIGHT, AMAR, WMAR, STATUS )

Arguments

SIGMA = REAL (Given)
Standard deviation of the Gaussian to be used for smoothing.
IBOX = INTEGER (Given)
Half-size, in pixels, of the box over which the Gaussian smoothing profile will be applied (the actual box used has an edge which is 2∗IBOX+1 pixels long). This defines the region within which the point spread function is non-zero.
SAMBAD = LOGICAL (Given)
If a .TRUE. value is given, then any " bad" pixels in the input array A will be propagated to the output array B (output values will be calculated for all other output pixels). If a .FALSE. value is given, then the WLIM argument is used to determine which output pixels will be bad (if any). This argument is not relevant if BAD is .FALSE..
WLIM = ? (Given)
The minimum weighting that can be used to compute a smoothed output pixel if SAMBAD is .FALSE.. Any output pixels falling short of the specified weight will be set to the bad value (invalid pixels carry no weight, others have Gaussian weights about the central pixel). The value must be greater than 0.0 and should be less than or equal to 1.0. This argument is not used if SAMBAD is .TRUE. or if BAD is .FALSE..
NX = INTEGER (Given)
First dimension of the two-dimensional array A.
NY = INTEGER (Given)
Second dimension of the two-dimensional array A.
BAD = LOGICAL (Given)
Whether or not it is necessary to check for bad pixels in the input array A.
VAR = LOGICAL (Given)
If a .FALSE. value is given for this argument, then smoothing will be performed as if the array supplied (A) is an array of data and the PSF will be used directly as specified. If a .TRUE. value is given, then smoothing will be applied as if the array is an array of variance values associated with an array of data; in this case, the effective PSF will be reduced in width by a factor 2 and the mean output values will be reduced to reflect the variance-reducing effect of smoothing.
A( NX, NY ) = ? (Given)
Array containing the input image to be smoothed.
B( NX, NY ) = ? (Returned)
Array to receive the smoothed output image.
BADOUT = LOGICAL (Returned)
Whether there are bad pixel values in the output array.
WEIGHT( 2 ∗ IBOX + 1 ) = ? (Returned)
Workspace for the Gaussian weighting function.
AMAR( NX ) = ? (Returned)
Workspace for the weighted sum of array values.
WMAR( NX ) = ? (Returned)
Workspace for the sum of pixel weights.
STATUS = INTEGER (Given and Returned)
The global status.

Notes:

There are routines for processing double precision and real data. Replace " x" in the routine name by D or R as appropriate. The data types of the WLIM, A, B, WEIGHT, AMAR, and WMAR arguments must match the routine used.