The first file is the input image file that is to be filtered. The second file, if present, contains the entire window of filter weights, in a VICAR labelled image file. If the second file is present, the parameters NLW, NSW, SYMM, and WEIGHTS are ignored, since their values are either supplied or implied by the contents of the secon input file.
A file to write the filtered product into
The standard size field defining the area of the input picture that is to be filtered.
Starting line of the area to be filtered.
Starting sample of the area to be filtered
Number of lines in the area to be filtered.
Number of samples in the area to be filtered.
This specifies the number of lines of weights. Must be odd.
This specifies the number of samples in each line of the weight matrix. Must be odd.
This keyword parameter indicates which portion of the weights matrix is
given in the WEIGHTS parameter. When only part of the weights matrix is
given, the remaining parts are generated by reflection through the center
lines.
UL - (the default) indicates that the upper left quadrant has
been specified.
UPPER - indicates that the upper half of the matrix has been given.
LEFT - indicates that the left half of the matrix has been given.
ALL - indicates that all elements of the matrix have been given.
For compatibility, the values NONSYMME and ASYMMETR are also allowed. They
have the same meanings as UPPER and ALL, respectively.
These are the weights to be used in the filter. They should be ordered from left to right, the top to bottom.
The result of the filtering and rescaling (via SCALE parameters) is divided
by this value prior to the clipping to the RANGE parameter.
Each output point O(l,s) is given by:
O(l,s) = GAIN*(T(l,s)/DIVIDE) + OFFSET
where GAIN and OFFSET are defined by SCALE and T is the output of the
convolution. The default is that the sum of the weights is used. If this
sum is zero, then 1.0 is used.
This keyword specifies the application of a linear transformation to each
filtered value T(l,s):
O(l,s) = GAIN*(T(l,s)/DIVIDE) + OFFSET
where SCALE=(OFFSET,GAIN). Default is SCALE=(0,1).
See also DIVIDE.
All pixels with DN less than DNMIN upon output are set to DNMIN. DNMIN is synonymous with RANGE(1). The default is the least value of the data format.
All pixels with DN greater than DNMAX upon output are set to DNMAX. DNMAX is synonymous with RANGE(2). The default is the largest value of the data format.
The pair of values (a,b) specifies the range of DN to which the output will be clipped. RANGE is synonymous with (DNMIN, DNMAX). If both RANGE and DNMIN or DNMAX are specified, then RANGE will take precedence.
Specifies image handling at image boundaries. Setting EDGE=CYCLE or 'CYCLE causes the program to treat the image as if it wrapped around at boundaries in both directions. 'LCYCLE and 'SCYCLE cause wrap-around in the line and sample direction only, respectively. The default is for the program to reflect the image at the boundaries.
set. Any combination of the allowable parameters may be given. If any of the parameters are given interactively, the interactive value takes precedence.