
	This directory contains input files for spectrum.
	Thanks to Jim White of Los Alamos National Laboratories for
	providing the pre-clustered images of Albuquerque.

albq1.xv -  An image of Albuquerque;  this pre-clustered VIFF image has 
	    additional map information derived from the clustering algorithm
	    used to generate it;  it has image data of type BYTE. 
	    There are 6 map columns, each of size 256, with values between 0
	    and 255.  It does not have a count or class column; spectrum
	    will generate the class column for it on input, and a class column
	    for it according to the class categories created on output.
	    There is no legend file to go with it, as it has no class column.
	    Note that because this image has FLOAT map data, it cannot be
	    used as input to putimage or editimage.  It has image->ispare1 = 0,
	    indicating that there is only mean vector information in the map
	    data (no class column, count column, covariance matrix, etc).
	    It has image->ispare1 = 6, indicating that there are 6 columns in
	    the map data.  It is typical of a pre-clustered image that might
	    be used as input to spectrum for classification.  For example:
	    % spectrum -i albq1.xv


albq2.xv -  An image of the Rio Grande River and the west side of Albq, in the
	    same format as albq1.xv.

albq3.xv -  An image of the far North East side of Albq and the Sandia 
	    Mountains, in the same format as albq1.xv and albq2.xv

ngl.xv   -  This is a "normal" VIFF image of a Near Ground Level plot in the
	    Sevilleta Wildlife Refuge, taken as part of the Long Term Ecological
	    Research Project currently underway at the Sevilleta, as well as at
	    other sites across the United States.  It was not pre-clustered,
	    and has no additional data stored in its map data.  It is typical
	    of a non pre-clustered image that might be used as input to spectrum
	    for classification.  For example:
	    % spectrum -i ngl.xv

ngl.classed - This is the image created by spectrum when run with ngl.xv as
	    an input file, and classified to have 3 class categories;  to
	    restore the complete legend, you must also use ngl.lgd.

ngl.lgd  -  This is the legend file that goes with ngl.classed.  To see the
	    classification of ngl.xv, you must run:
	    % spectrum -i ngl.classed -l ngl.lgd
