Help for IMGSTAT

PURPOSE

	IMGSTAT is a VICAR applications program for calculating statistical
	quantities in local areas surrounding each pixel in an input image.
    It is a sizing down operation producing images zoomed down by the
    window size.  
	The local mean, minimum, maximum, standard deviations, and optionally
        the mean slope devided by the standard deviation are quantities that 
        are obtained.

EXECUTION FORMAT

	TAE>IMGSTAT IN,OUT, WINDOW

	where

	IN	is the input data set (VICAR labeled image).

	OUT	is the output image of the selected statistic.
                Three output files are required; optionally,
                if four output files are requested, a standard
                deviations output file is created.

	WINDOW	is the number of lines in the local area window.
OPERATION

	IMGSTAT performs a sliding window statistical analysis on an input
	image. An output pixel's position represents the center position of
	the window for the input image and its value represents statistics
	based on data within the window only. The window moves along one sample
	at a time and one line at a time until each pixel in the input has
	been at the center of the window. In other words statistics are 
	compiled for the local area around each and every pixel in the input
	image.

	The edge conditions are handled as follows. Any window positions that
	extend beyond the edge of the image are handled by using data from the
	nearest pixel on the image. This is not a reflection scheme like some
	sliding window algorithms have.


    Byte, half, full and real formats are supported. The output
    format is the same as whatever is input. The internal calculations
    are performed in real format. The vicar xv routines handles the
    output conversion.
	Input images must have less than 14001 samples. There is no limit 
    to the number of lines in the input.
EQUATIONS

	MIN
		The min is a smallest value of all DN values in the 
		window.

	MAX
		The max is a largest value of all DN values in the 
		window.

	MEAN
		The mean is a result of the sum of all DN values in the 
		window divided by the number of pixels in the window.
		(the average of the DNs)
			WINDOWSUM / NPIXWNDO

    STANDARD DEVIAION
        The standard deviation is the square root of the sum 
        of all observations minus the mean devided by the 
        number of observations.     


     MEAN SLOPE / STANDARD DEVIAION
        The mean slope is computed and is then devided by the 
        standard deviaion for all observations.
  

Original Programmer:	Barbara McGuffie    February 1995
Cognizant Programmer:	Ray Bambery     

History:
        2/95  - B. McGuffie  
        3/96  - B. McGuffie     Increased samples to 14000
        10/96 - B. McGuffie
        1/97  - N. Ritter       Added Slope deviation output options.
        13-Apr-2001 - R Bambery - Converted to real*4 internals.
                                Now allow BYTE, HALF, FULL and REAL formats