20001130: IMGREMAP: MAG= vs RES= (cont.)

>From: Jason Allard <address@hidden>
>Organization: PSU
>Keywords: 200010240345.e9O3jC425238 McIDAS-X IMGREMAP MAG RES

Jason,

>I'm using your suggestion of the EB command to stretch the gray scale of the
>image, however, I'm having some difficulty.  When all I have is 0s and 1s
>in the array, the command:
>
>EB 0 1 0 255
>
>turns the entire screen white.

The problem you are going to have to deal with if the values in your
image are just 0 or 1 is the grouping of display values by the number
of display "colors" (display levels) that you are running McIDAS in.
For instance, if you start McIDAS and say that you want to use 64 shades
of gray for imagery (this would be done by specifying '-imageColors 64'
as an option on the McIDAS invocation command line), then groups of 4
pixels would be bined together in the same output "color":

0 - 3 -> display level 0
4 - 7 -> display level 1
 etc.

If you specified 128 shades of gray, then every two imagery levels would
be bined together into a single display level:

0 - 1 -> display level 0
2 - 3 -> display level 1
 etc.

So, if you really are dealing with only 0 and 1 values, you have to
stretch them into a different display range.  You can do this using
the SU facility.

>I thought that it would keep the 0s black
>and the 1s white.  When I have 0s and 2s (ran the test twice with the same
>thresholds, the command:
>
>EB 0 2 0 255 turns the screen gray and the 2s are a lighter gray.
>How would I keep the screen black, i.e., the 0s black.

The explanation above should help you understand what is going on.  At
this point, you would change the enhancement associated with the display:

EU MAKE 0 127 BLACK
EU MAKE 128 255 WHITE

If the EB command mapped non-zero data values into display levels between
128 and 255, then the combination of the EU MAKE commands should do
what you want.  If there is a range of non-zero values, then you will
want to find out what the distribution of these values is.  The ASTAT
command can do this for you easily.  After you know what the distribution
of values is, you will be better able to come up with a stretch of
data values to display values and colorization that helps you visualize
the information in the image.

I must add, however, that NACOMP was most likely written with the intention
of then running the PERCENT option to get some numbers, not to produce
an image that would be visually appealing.

Tom