Level 2 Help for MEDIAN

INP

STRING - INP=A where A is the input dataset name.


OUT

STRING - OUT=B where B is the output dataset name.


SIZE

4 INTEGERS - SIZE=(SL,SS,NL,NS) where SL is the starting line, SS is the
 starting sample, NL is the number of lines output and NS is the number of 
 samples output. (SIZE is usually defined as SIZE=(1,1,NL,NS)). Default is 
 taken from the VICAR label within the program.


NL

INTEGER - NL=N1 where is N1 is the number of lines output.


NS

INTEGER - NS=N1 where is N1 is the number of samples output.


NLW

INTEGER - NLW=I1 where I1 is an integer and specifies the size of the filter
 kernel in lines. NLW should be an odd integer; if it is not, the program
 uses the next larger odd integer.  Default is NLW=11. 


NSW

INTEGER - NSW=I2 where I2 is an integer and specifies the size of the filter
 kernel in samples. NSW should be an odd integer; if it is not, the program
 uses the next larger odd integer.  Default is NSW=11.


HIGHPASS

KEYWORD - Valid:('HIGH) 'HIGHPASS specifies that the output is to be in
 highpass format, i.e., the input minus the local median value. Default
 is the lowpass mode.


DCLEVEL

DCLEVEL is a constant added to all pixels prior to output, when in the 
HIGHPASS mode.  It is most commonly used to rescale byte data back into
the 0 to 255 range.


PERCENT

REAL - PERCENT=R4 where R4 is an floating point number and specifies the
 percentage of the filter window values which are less than or equal to the
 output value.  The default of PERCENT=50.0 specifies that the true median 
 value of filter window is to be found, whereas if a lower (or higher)
 value is specified, the output value will be somewhat lower (or higher)
 than the actual median.


DCTRAN

DCTRAN is a scaling factor, used in the HIGHPASS mode, when the user desires
to add back part of the input image to the output. In the HIGHPASS mode

      OUTPUT  =   INPUT - MEDIAN + (DCTRAN*INPUT) + DCLEVEL