This routine calculates the Laplacian of the supplied two-dimensional NDF , and subtracts it from the original array to create the output NDF. The subtractions can be done a specified integer number of times. This operation can be approximated by a convolution with the kernel:

$\begin{array}{ccc}\hfill N\hfill & \hfill -N\hfill & \hfill -N\hfill \\ \hfill N\hfill & \hfill +8N\hfill & \hfill -N\hfill \\ \hfill N\hfill & \hfill -N\hfill & \hfill -N\hfill \end{array}$

where N is the integer number of times the Laplacian is subtracted. This convolution is used as a uni-directional edge detector. Areas where the input data array is flat become zero in the output data array.

IN = NDF (Read)

NUMBER = _INTEGER (Read)

OUT = NDF (Write)

TITLE = LITERAL (Read)

laplace a 10 b

This subtracts ten Laplacians from the NDF called a, to make the NDF called b. NDF b inherits its title from a.

KAPPA: SHADOW, MEDIAN; FIGARO: ICONV3.

• This routine correctly processes the WCS, AXIS, DATA, and VARIANCE components of an NDF data structure. QUALITY is propagated.

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

• All non-complex numeric data types can be handled.