.INPUT FILE: PIC001
[Enter the name of the file to be interpolated.]
.OUTPUT FILE: IMG024
[Enter the file where the interpolated image is to be placed.]
For 2-D input image:
.ENTER DIMENSIONS, NSAM & NROW: 128,128
[Enter the dimensions for the interpolated image.]
For 3-D input image:
.ENTER DIMENSIONS, NSAM, NROW, & NSLICE: 128,128,128
[Enter the dimensions for the interpolated image.]
NOTES
This operation can be used for either interpolation or
extrapolation.
The output image does not have to have the same
proportions (ratio NSAM/NROW) as the input picture.
Only the new X dimension
(NSAM) of the output image need be entered.
The other dimension(s) will then be computed
assuming the same proportions as the input image. For
example, let the input image have the size 60x40.
Then the sequence
.ENTER DIMENSIONS, NSAM & NROW: 120
will have the same effect as:
.ENTER DIMENSIONS, NSAM & NROW: 120,80
The program works by throwing a new grid onto the image
whose boundaries coincide with those of the old grid. Each
pixel of the new grid is computed from its old neighbors by
bilinear interpolation or triangular interpolation.
Note that for raw data, a reduction in scale by 'IP'
results in a loss of S/N ratio, since only a subset of the
original points may be used for interpolation. The proper
procedure maintaining the original information as much as
possible is to apply an appropriate low-pass filtration to
the original image (either 'FQ', or for large images design s
Remez filter using 'FF PLOT' and use 'RC'), and decimate it
using 'DC' operation.