coyote: CGIMAGE

Description
The purpose of this program is to create a TV command that works the way
the TV command would be expected to work if it was written today, rather
than 25 years ago. In other words, it knows the difference between an
8-bit device and a 24-bit device, it honors the POSITION keyword like
other graphics commands in IDL, it honors the !P.MULTI value, like other
graphics commands in IDL, it works seamlessly with both 8-bit and 24-bit
images. In addition to other modern features, this program can also
display images that contain an alpha channel and can display transparent
images.
There is a common block in cgImage that is defined as follows upon exiting
this command::
   COMMON FSC_$CGIMAGE, $
      _cgimage_xsize, $    ; The X size of the image.
      _cgimage_ysize, $    ; The Y size of the imge.
      _cgimage_winxsize, $ ; The X size of the window displaying the image.
      _cgimage_winysize, $ ; The Y size of the window displaying the image.
      _cgimage_position, $ ; The final position of the image in the window.
      _cgimage_winID, $    ; The window index number of the window displaying the image.
      _cgimage_current     ; Set to 1 if a call to cgImage is made.
The program requires the `Coyote Library `
to be installed on your machine.
To learn more about transparent images and cgImage see the article
`Working with Transparent Images and cgImage `.
Categories
Graphics
Examples
To display an image with a contour plot on top of it, type::
   filename = FILEPATH(SUBDIR=['examples','data'], 'worldelv.dat')
   image = BYTARR(360,360)
   OPENR, lun, filename, /GET_LUN
   READU, lun, image
   FREE_LUN, lun
   thisPostion = [0.1, 0.1, 0.9, 0.9]
   cgIMAGE, image, POSITION=thisPosition, /KEEP_ASPECT_RATIO
   CONTOUR, image, POSITION=thisPosition, /NOERASE, XSTYLE=1, $
       YSTYLE=1, XRANGE=[0,360], YRANGE=[0,360], NLEVELS=10
To display four images in a window without spacing between them::
   !P.Multi=[0,2,2]
   cgImage, image, Margin=0
   cgImage, image, Margin=0
   cgImage, image, Margin=0
   cgImage, image, Margin=0
   !P.Multi = 0
To display four image in a window with associated color bars::
   !P.Multi=[0,2,2]
   p = [0.02, 0.3, 0.98, 0.98]
   LoadCT, 0
   cgImage, image, Position=p
   cgColorbar, Position=[p[0], p[1]-0.1, p[2], p[1]-0.05]
   p = [0.02, 0.3, 0.98, 0.98]
   LoadCT, 2
   cgImage, image, Position=p
   cgColorbar, Position=[p[0], p[1]-0.1, p[2], p[1]-0.05]
   p = [0.02, 0.3, 0.98, 0.98]
   LoadCT, 3
   cgImage, image, Position=p
   cgColorbar, Position=[p[0], p[1]-0.1, p[2], p[1]-0.05]
   p = [0.02, 0.3, 0.98, 0.98]
   LoadCT, 5
   cgImage, image, Position=p
   cgColorbar, Position=[p[0], p[1]-0.1, p[2], p[1]-0.05]
   !P.Multi =0
To set a missing value to -32767 and the color white and do
a 3% histogram clip of the image::
   cgLoadCT, 4, /Brewer, NColors=254
   TVLCT, palette, /Get
   cgImage, image, Missing_Value=-32767, Missing_Color='white', Stretch='Clip', Clip=3
To display a transparent image on top of another image::
   cgImage, cgDemoData(7), CTIndex=0
   cgImage, cgDemoData(5), CTIndex=33, Transparent=50, $
        AlphaFGPosition=[0.25, 0.25, 0.75, 0.75], Missing_Value=0
mage:: cgimage.png
Author
FANNING SOFTWARE CONSULTING::
    David W. Fanning
    1645 Sheely Drive
    Fort Collins, CO 80526 USA
    Phone: 970-221-0438
    E-mail: david@idlcoyote.com
    Coyote's Guide to IDL Programming: http://www.idlcoyote.com/
History
Modification History::
  Written by: David W. Fanning, from modifications to TVIMAGE. 3 Feb 2011.
  8 Feb 2011. Added OPOSITION keyword. DWF.
  27 Feb 2011. Added keywords to make cgImage more compatible with TVImage calls. DWF.
  Color table vectors must be obtained AFTER loading the color palette. 6 March 2011. DWF.
  I have been convinced (conversations with Wayne Landsman) that if the
     CENTER keyword is set, the MINUS_ONE keyword is not needed, since
     it was created to solve the same problem. So, I have changed the
     default setting of MINUS_ONE to 0. 14 March 2011. DWF.
   Corrected a problem with restoring color tables if a PALETTE is used. 31 March 2011. DWF.
   Whoops! Documented a CHARSIZE keyword, but forgot to define it. 7 July 2011.
   Damnation! I did the same thing with the FONT keyword! 25 July 2011.
   And now a TITLE keyword! What the devil is going on!? 29 Aug 2011.
   Very slight modifications to image size and start position so that the image is
      positioned completely inside the axes. 30 Sept 2011. DWF.
   Fitting the image inside the axes causes image matching problems (and lines!) in
      other programs, so I've decided to only do positioning inside axes when the
      user asks for this capability by setting the new FIT_INSIDE keyword. 16 Nov 2011. DWF.
   Problem with transparent images with alpha channels caused by changes in the TVImage->cgImage
      transition. Added AlphaFGPosition keyword to address issues. Cleaned up the
      code and improved the internal documentation. 22 Nov 2011. DWF.
   Added the ability to stretch 2D image arrays in various ways before display. 1 Dec 2011.
   Added the ability to handle missing data in 2D arrays before display. 1 Dec 2011.
   Added a DISPLAY keyword to display the image in windows with the image aspect ratio. 2 Dec 2011.
   Added the ability to send the output directly to a file via the OUTPUT keyword. 9 Dec 2011, DWF.
   PostScript, PDF, and Imagemagick parameters can now be tailored with cgWindow_SetDefs. 14 Dec 2001. DWF.
   Modified to use cgDefaultColor for default color selection. 24 Dec 2011. DWF.
   Changes to allow better default colors, based on changes to cgColor and cgDefaultColor. 1 Feb 2012. DWF.
   I was passing the wrong MINVALUE and MAXVALUE values to the scaling function. 12 Feb 2012. DWF.
   I made cgImage aware of a "feature" of MAP_SET that sets !P.MULTI[0]=-1, which was screwing
      up the algorithm to cause cgImage to erase the display window. 28 Feb 2012. DWF.
   Added a Standard Deviation stretch, including the EXCLUDE and MULTIPLIER keywords to the
      SDevScl command. 6 June 2012. DWF.
   Now saving the image POSITION in FSC_$CGIMAGE common block, even if in PostScript, because other
      routines (e.g., cgMap) may depend on it (e.g., using ONIMAGE keyword). 26 July 2012. DWF.
   Added the ability to use escape characters in plot titles to specify cgSymbol symbols. 27 July 2012. DWF.
   Modified the way the XRANGE and YRANGE keywords work when the OVERPLOT keyword is also set. In this case,
      I will modify the image position to honor the XRANGE and YRANGE values with respect to the axes that the
      image is being overplot onto. 15 August 2012. DWF.
   Set the NOERASE keyword if the OVERPLOT keyword is set and NOERASE is undefined. 16 Aug 2012. DWF.
   Modified the way the HIST_EQUAL stretch works. Previously the image was displayed as all zeros
      if the input image had a minimum value less that zero. 21 Aug 2012. DWF.
   Fixed a problem with the INTERPOLATE variable that prevented interpolaton from occurring. 12 Sept 2012. DWF.
   A misspelled MULTIPLIER keyword was interfering with a Standard Deviation stretch. 16 Sept 2012. DWF
   Added the TRANSPARENT keyword to allow transparent display of images. 17 October 2012. DWF.
   Added the MAPCOORD keyword to allow the XRANGE and YRANGE of the image to be specified by the map
      coordinate object. 17 October 2012. DWF.
   Added CTINDEX, BREWER, and REVERSE keywords to make loading a color table palette easier. 17 October 2012. DWF.
   Now setting MISSING_VALUE pixels to completely transparent if a transparent image is created. 17 October 2012. DWF.
   Added the ability to apply a stretch to a 2D image prior to converting it to a transparent image. 18 October 2012.DWF.
   Added a FILENAME keyword so that files of known format (GeoTIFF, JPEG, PNG, etc.) can be read to supply an
      image for display. 18 October 2012. DWF.
   Fixed a problem that prevented transparent images from be displayed with !P.Multi. 20 Oct 2012. DWF.
   Provided a fix to allow transparent images in versions of IDL from IDL 6.4 and earlier. 18 Nov 2012. DWF.
   When reading a GeoTiff file, the map object created should be named mapCoord, not map, so that
      the image data ranges can be set automatically. 11 January 2013. DWF.
   Additional work to allow overplotting of transparent images. 12 Jan 2013. DWF.
   For some reason I thought I had to have transparency ON when overplotting. This turns out to be wrong
       and gets in the way of outputting to PostScript and working with some map projected images. I've
       removed this requirement in the code. I also fixed a problem in which the OUTPUT position changes
       if the TRANSPARENT keyword is used. 27 January 2013. DWF.
   I had some code dealing with XRANGE and YRANGE properties of the image when a map coodinate object
       was used with cgImage. This was inadvertently setting the XRANGE and YRANGE properties of the
       image, which was in turn causing the range values to be SAVED in the plotting system variables.
       This interferred with backward compatibility with the TV command, so I have removed it. 31 Jan 2013. DWF.
   Whoops! Typo in my last fix. Getting too old, I guess. 6 Feb 2013. DWF.
   Setting any of the MISSING_*** keywords while doing multiple plots resulted in the value
       of !P.Multi being ignored for the image. This is fixed for now, but just a warning. Setting
       these keywords creates transparent images, and makes things MUCH more complicated. So, I'm
       probably at the limit of what is possible now. :-) 30 April 2013. DWF.
   The LAYOUT keyword went on walkabout after the last changes. Restored to operation. 12 July 2013. DWF.
   The YTITLE keyword was missing when passed to cgWindow. Fixed now. 24 Oct 2013. DWF.
   Fixed a couple of places where I meant "missing_index" and used "missing_color". 26 Jan 2014. DWF.
   Added check for open graphics window when displaying alpha-channel image. 31 March 2014. DWF.
   Added XVECTOR and YVECTOR keywords. 1 April 2014. DWF.
   Fixed a problem in which the POSITION of the image was specified as an integer array when it
        should have been a floating point array. 8 January 2015. DWF.
   Added compression stretch and updated retired program references. 27 Mar 2015. DWF.
Copyright
Copyright (c) 2011-2015, Fanning Software Consulting, Inc.
 function creates a transparent image out of a normal 2D or 3D image.
Params
image:  in, required, type=various
    An 8-bit (MxN), 24-bit (e.g., MxNx3), or a 24-bit + alpha channel
    (e.g., MxNx4) image  to display.
x: in, optional, type=integer
    The X position of the lower-left corner of the image in device
    coordinates. This parameter is only recognized if the TV keyword
    is set. If the Y position is not used, X is taken to be the image
    "position" in the window. See the TV command documenation for details.
y: in, optional, type=integer
    The Y position of the lower-left corner of the image in device
    coordinates. This parameter is only recognized if the TV keyword
    is set.
Keywords
addcmd: in, optional, type=boolean, default=0
     Set this keyword to add this command to an already open cgWindow to
     be executed when the window is resized. If the DISPLAY keyword is also
     set, this keyword will act as if the WINDOW keyword were set.
alphabackgroundimage: in, optional, type=varies
     Normally, when a image with an alpha channel is displayed, the image is
     blended with whatever is currently in the display window. This means, the
     program has to obtain that background image. This is not a problem on devices
     (e.g., WIN, X, Z) that allow this kind of operation, but it is on devices
     (e.g., the PostScript device, PS) that do not. To get around this problem,
     you can pass the background image to the cgImage program. This background image
     will be blended with the alpha channel image you wish to display. If an alpha
     channel image is displayed on a device in which there is no way to obtain the
     background image, and this keyword is not used to pass a background image, then
     the alpha channel image will be blended with a white background image.
     This keyword is only used if an alpha channel image is passed to the
     program via the IMAGE parameter. The AlphaBackgroundImage does not need
     to have the same dimensions as the alpha channel image. The background image
     can be either a 2D image or a 24-bit image.
alphabgposition: in, optional, type=fltarr(4)
     The normalized position or portion of the background image that is going to be used
     to create the alpha channel image. Normally, and by default, the alphabgposition encompasses
     the entire graphics window, [0.0, 0.0, 1.0, 1.0].
alphafgposition: in, optional, type=fltarr(4)
     The normalized position in the background image where the input image is to be located.
     By default, the input image takes up the entire extent of the background image, [0.0, 0.0, 1.0, 1.0].
axis: in, optional, type=boolean, default=0
     A misspelled version of the AXES keyword. Provided as a service to people whose
     fingers have minds of their own.
axes: in, optional, type=boolean, default=0
     Set this keyword to display the image with axes surrounding the image. If the POSITION
     keyword is not used, a MARGIN of 0.1 is used to allow the axes to show. If you wish
     to make the image fit entirely inside the axes, set the `FIT_INSIDE` keyword. Otherwise,
     the axis sit on top of the image data.
axkeywords: in, optional, type=structure
     A structure of AXIS keywords and values that can be used to configure the axes
     in whatever way the user desires. Many of the most often used axis keywords are available
     as cgImage keywords. For example::
        IDL> axis_format = {XTicks:4, XTickname:['Cow', 'Pig', 'Dog', 'Cat', 'Owl']}
        IDL> cgImage, cgDemoData(7), AXKEYWORDS=axis_format, /Axes, XRange=[0,20]
background: in, optional, type=string, default='white'
     The name of the background color for the image display. Unlike the TV command in IDL,
     the cgImage command will erase the display before executing the command like other
     fundamental graphics commands (e.g., Plot, Contour, Surface) in IDL.
beta: in, optional, type=float, default=3.0
     The beta factor in a Hyperpolic Sine stretch. Available only with 2D images.
bottom: in, optional, type=integer, default=0
     If the SCALE keyword is set, the image is scaled before display so that all
     displayed pixels have values greater than or equal to BOTTOM and less than
     or equal to TOP. Available only with 2D images.
brewer: in, optional, type=boolean, default=0
     This keyword is used only if the `CTIndex` keyword is used to select a color table number.
     Setting this keyword allows Brewer color tables to be used.
charsize: in, optional, type=float
     Sets the character size. Used only if the AXES keyword is also set. By default,
     the value from cgDefCharsize().
clip: in, optional, type=float, default=2
     A number between 0 and 50 that indicates the percentage of pixels to clip
     off either end of the image histogram before performing a linear stretch.
     Available only with 2D images.
color: in, optional, type=string, default='opposite'
     The name of the color in which to draw the axes. Used only if the `AXES` keyword is set.
ctindex: in, optional, type=integer
     The index number of a color table. The `Brewer` and `Reverse` keywords will be checked
     to see how to load the color table into the `Palette` keyword. This keyword will take
     precidence over any colors that are loaded with the `Palette` keyword. The default is
     to use whatever colors are loaded in the current hardware color table.
display: in, optional, type=boolean, default=0
     If this keyword is set, a new display window is created (with cgDisplay) that has the
     same aspect ratio as the image. The image is displayed in that window. If the WINDOW
     keyword is also set, a new cgWindow is created with the WASPECT keyword set to the image
     aspect ratio, and the image is displayed in that new cgwindow.
erase: in, optional, type=boolean, default=1
     An obsolete keyword. Used only for compatibility with the earlier TVImage command. The
     default for cgImage is to always erase the graphics display before displaying the image
     unless told otherwise by setting the `NOERASE` keyword. This makes cgImage consistent with
     other IDL graphics commands.
exclude: in, optional
     The value to exclude in a standard deviation stretch.
exponent: in, optional, type=float, default=4.0
     The logarithm exponent in a logarithmic stretch. Available only with 2D images.
filename: in, optional, type=string
     The name of a file that IDL can read with READ_IMAGE (e.g, GEOTIFF, TIF, JPEG, PNG, etc.).
     The file is read to obtain the image to be displayed.
fit_inside: in, optional, type=boolean, default=0
     When the AXES keyword is set, the default is to position the axes on top of the image
     using the POSITION. However, if this keyword is set, the axes are positioned at POSITION
     and the image is sized so that it fits entirely inside the axes.
font: in, optional, type=integer
     This keyword selects the font used for axis and title display. The default is to use
     the value of !P.Font.
gamma: in, optional, type=float, default=1.5
     The gamma factor in a gamma stretch. Available only with 2D images.
interpolate: in, optional, type=boolean, default=0
     Set this keyword to interpolate with bilinear interpolation the display image as it
     is sized to its final position in the display window. Interpolation will potentially
     create image values that do not exist in the original image. The default is to do no
     interpolation, so that image values to not change upon resizing. Interpolation can
     result in smoother looking final images.
keep_aspect_ratio: in, optional, type=boolean, default=0
     By default, the output image is resized into the `POSITION` in the graphics window.
     This can result in a distortion of the image aspect ratio (the Y size of the image
     divided by the X size of the image). Setting this keyword will preserve the original
     aspect ratio of the image in the output display window. In effect, the image will
     be placed in the window `POSITION` in a way that preserves its aspect ratio. The
     actual final location of the image in the display window can be obtained via the
     `OPOSITION` keyword.
layout: in, optional, type=intarr
     This keyword specifies a grid with a graphics window and determines
     where the graphic should appear. The syntax of LAYOUT is a 3-element
     array: [ncolumns, nrows, location]. The grid is determined by the
     number of columns (ncolumns) by the number of rows (nrows). The location
     of the graphic is determined by the third number. The grid numbering
     starts in the upper left (1) and goes sequentually by column and then
     by row. Note that using the LAYOUT keyword automatically sets the NOERASE
     keyword to 1.
mapcoord: in, optional, type=object
     An object reference to a map coordinate object (e.g., cgMap). If present, the
     XRANGE and YRANGE keyword values will be obtained from this map object, if they
     are not defined otherwise.
margin: in, optional, type=float, default=0.0
     A single value, expressed as a normalized coordinate, that
     can easily be used to calculate a position in the window.
     The margin is used to calculate a `POSITION` that gives
     the image an equal margin around the edge of the window.
     The margin must be a number in the range 0.0 to 0.333. This
     keyword is ignored if the `POSITION` or `OVERPLOT` keywords are
     used. It is also ignored when cgImage is executed in a
     multi-plot window, EXCEPT if it's value is zero. In this
     special case, the image will be drawn into its position in
     the multi-plot window with no margins whatsoever. (The
     default is to have a slight margin about the image to separate
     it from other images or graphics. The default margin is 0.05.)
maxvalue: in, optional, type=varies
     If this value is defined, the data is linearly scaled between MINVALUE
     and MAXVALUE. MAXVALUE is set to MAX(image) by default. Setting this
     keyword to a value automatically sets `SCALE` to 1. If the maximum value of the
     image is greater than 255, this keyword is defined and SCALE=1.
mean: in, optional, type=float, default=0.5
     The mean factor in a logarithmic stretch. Available only with 2D images.
minus_one: in, optional, type=boolean, default=0
     The value of this keyword is passed along to the cgResizeImage
     command. It prevents cgResizeImage from adding an extra row and
     column to the resulting array, which can be a problem with
     small image arrays.
minvalue: in, optional, type=varies
     If this value is defined, the data is linearly scaled between MINVALUE
     and `MAXVALUE`. MINVALUE is set to MIN(image) by default. Setting this
     keyword to a value automatically sets SCALE=1. If the minimum value of the
     image is less than 0, this keyword is defined and SCALE=1.
missing_color: in, optional, type=string, default='white'
     The color name of the missing value. Available only with 2D images.
missing_index: in, optional, type=integer, default=255
     The index of the missing color in the final byte scaled image. Available only with 2D images.
missing_value: in, optional, type=integer
     The number that represents the missing value in the image. Available only with 2D images.
multimargin: in, optional, type=varies
     Sometimes, when displaying multiple images with !P.Multi, you
     want the images to be slightly smaller than the position created
     by !P.Multi so you can add, for example, a colorbar or an annotation
     to the image. This keyword can be used to adjust the image position
     by a small margin. A four-element array, the margins apply to the
     [bottom, left, top, right] of the image position. So, to
     leave room at the top of an image for a color bar, you might
     type this::
         cgImage, image, MultiMargin=[0, 0, 4, 0]
     This keyword applies *only* to images displayed with !P.Multi, and if
     passed a scalar value, will use the same value for all four positions.
multiplier: in, optional, type=float
     The multiplication factor in a standard deviation stretch. The standard deviation
     is multiplied by this factor to produce the thresholds for a linear stretch.
ncolors: in, optional, type=integer, default=256
     If this keyword is supplied, the `TOP` keyword is ignored and the TOP keyword
     is set equal to  NCOLORS-1. This keyword is provided to make cgImgScl easier
     to use with the color-loading programs such as cgLOADCT::
          cgLoadCT, 5, NColors=100, Bottom=100
          scaled = cgImgScl(image, NColors=100, Bottom=100)
     Setting this keyword to a value automatically sets SCALE=1 and STRETCH='LINEAR', if not
     defined otherwise. Available only with 2D images.
negative: in, optional, type=boolean, default=0
     Set this keyword if you want to display the image with a negative or reverse stretch.
     Available only with 2D images.
noerase: in, optional, type=boolean, default=0
     Set this keyword to prevent the command from first erasing the graphics
     display before displaying the image.
nointerpolation: in, optional, type=boolean, default=0
     This is an obsolete keyword that is included here only for compatibility with
     the older TVImage command. Use the `INTERPOLATE` keyword.
normal: in, optional, type=boolean, default=0
     Setting this keyword means image position coordinates x and y are interpreted
     as being in normalized coordinates. This keyword is only valid if the TV
     keyword is set.
oposition: out, optional, type=float
     Set this keyword to a named variable to obtain the actual position in the
     graphics window where the image was located. The output position may be
     different from the input `POSITION`, especially if the `KEEP_ASPECT_RATIO`
     keyword is set. Note that the output position is also stored in a
     FSC_$CGIMAGE common block so that other programs (e.g., cgMap, cgImageInfo,
     etc.) can take advantage of this information.
outfilename: in, optional, type=string
     If the `Output` keyword is set, the user will be asked to supply an output
     filename, unless this keyword is set to a non-null string. In that case, the
     value of this keyword will be used as the filename and there will be no dialog
     presented to the user.
output: in, optional, type=string, default=""
     Set this keyword to the type of output desired. Possible values are these::
        'PS'   - PostScript file
        'EPS'  - Encapsulated PostScript file
        'PDF'  - PDF file
        'BMP'  - BMP raster file
        'GIF'  - GIF raster file
        'JPEG' - JPEG raster file
        'PNG'  - PNG raster file
        'TIFF' - TIFF raster file
     Or, you can simply set this keyword to the name of the output file, and the type of
     file desired will be determined by the file extension. If you use this option, the
     user will not be prompted to supply the name of the output file.
     All raster file output is created through PostScript intermediate files (the
     PostScript files will be deleted), so ImageMagick and Ghostview MUST be installed
     to produce anything other than PostScript output. (See cgPS2PDF and cgPS_Close for
     details.) And also note that you should NOT use this keyword when doing multiple
     plots. The keyword is to be used as a convenient way to get PostScript or raster
     output for a single graphics command. Output parameters can be set with cgWindow_SetDefs.
overplot: in, optional, type=boolean, default=0
     Setting this keyword causes the POSITION keyword to be ignored
     and the image is positioned in the location established by the
     last graphics command. If the XRange and YRange keywords are also
     used, the image position is adjusted with respect to the current axes
     range.  Setting this keyword also sets the NoErase keyword, if it is not currently set.
     It is assumed that the overplotted images fits entirely inside or is coincident with the
     position or range of the previous image. If not, an informational error is generated and
     the position is forced into the range of 0 to 1.
palette: in, optional, type=byte
     Set this keyword to a 3x256 or 256x3 byte array containing the RGB color
     vectors to be loaded before the image is displayed. Such vectors can be
     obtained, for example, from cgLoadCT with the RGB_TABLE keyword::
            cgLoadCT, 4, /BREWER, /REVERSE, RGB_TABLE=palette
            cgImage, cgDemoData(7), PALETTE=palette
position: in, optional, type=float
     The location of the image in the output window. This is a four-element
     floating array of normalized coordinates of the type given by !P.POSITION
     or the POSITION keyword to other IDL graphics commands. The form is [x0, y0, x1, y1].
     The default is [0.0, 0.0, 1.0, 1.0]. Note that this keyword is ALSO an output
     keyword. That is to say, upon return from cgImage this keyword (if passed by
     reference) contains the actual position in the window where the image was
     displayed. This may be different from the input values if the KEEP_ASPECT_RATIO
     keyword is set, or if you are using cgImage with the POSITION keyword when !P.MULTI
     is set to something other than a single plot. Note that the POSITION keyword should
     not, normally, be used when displaying multiple images with !P.MULTI. If it is used,
     its meaning differs slightly from its normal meaning. !P.MULTI is responsible for
     calculating the position of graphics in the display window. Normally, it would be a
     mistake to use a POSITION graphics keyword on a graphics command that was being drawn with
     !P.MULTI. But in this special case, cgImage will use the POSITION coordinates to calculate
     an image position in the actual position calculated for the image by !P.MULTI. The main
     purpose of this functionality is to allow the user to display images along with
     color bars when using !P.MULTI.
quiet: in, optional, type=boolean, default=0
     There are situations when you would prefer that cgIMAGE does not advertise itself by
     filling out the FSC_$CGIMAGE common block. For example, if you are using cgImage to
     draw a color bar, it would not be necessary. Setting this keyword means that cgImage
     just goes quietly about it's business without bothering anyone else.
reverse: in, optional, type=boolean, default=0
     Set this keyword to reverse the color table vectors selected with the `CTIndex` keyword.
save: in, optional, type=boolean, default=0
     Set this to cause a data coordinate system to be established for the image. The XRANGE
     and YRANGE keyword values will be used to establish a data coordinate system coincident
     with the final image position. Setting the `AXES` keyword automatically sets SAVE=1.
scale: in, optional, type=boolean, default=0
     Set this keyword to byte scale the image before display. If this keyword is not set,
     the image is not scaled before display. This keyword will be set automatically by using
     any of the keywords normally associated with byte scaling an image. Available only with
     2D images. If set, STRETCH is set to 1, unless it is set to another value.
stretch: in, optional, type=integer/string, default=0
     The type of scaling performed prior to display. May be specified as a number
     or as a string (e.g, 3 or "Log"). Available only with 2D images. If  Min(image)
     is less than zero or Max(image) is greater than 255, then the default value for
     stretch is 1.
     Number   Type of Stretch::
         0         None           No scaling whatsoever is done.
         1         Linear         scaled = BytScl(image, MIN=minValue, MAX=maxValue)
         2         Clip           A histogram stretch, with a percentage of pixels clipped at both the top and bottom
         3         Gamma          scaled = cgGmaScl(image, MIN=minValue, MAX=maxValue, Gamma=gamma)
         4         Log            scaled = LogScl(image, MIN=minValue, MAX=maxValue, Mean=mean, Exponent=exponent)
         5         Asinh          scaled = AsinhScl(image, MIN=minValue, MAX=maxValue, Beta=beta)
         6         SquareRoot     A linear stretch of the square root histogram of the image values.
         7         Equalization   A linear stretch of the histogram equalized image histogram.
         8         Gaussian       A Gaussian normal function is applied to the image histogram.
         9         MODIS          Scaling done in the differential manner of the MODIS Rapid Response Team
                                  and implemented in the Coyote Library routine ScaleModis.
         10        StdDev         Standard deviation stretch. scaled = SDevScl(image, MULTIPLIER=2).
sigma: in, optional, type=float, default=1.0
     The sigma scale factor in a Gaussian stretch. Available only with 2D images.
title: in, optional, type=string, default=""
     Set this keyword to the title of the image plot. No title will appear with the
     image unless the `AXES` keyword is also set.
top: in, optional, type=integer, default=255
     If the SCALE keyword is set, the image is scaled before display so that all
     displayed pixels have values greater than or equal to BOTTOM and less than
     or equal to TOP. Available only with 2D images.
transparent: in, optional, type=integer, default=50
     A number between 0 and 100 that specifies the percent of transparency between the
     image being displayed and the background image. Displaying a transparent image does
     not advance !P.Multi.
tv: in, optional, type=boolean, default=0
     Setting this keyword makes the cgImage command work much like the brain-dead
     TV command except that it will get colors right on all output devices. Most of
     the cgImage keywords are ignored if this keyword is set.
window: in, optional, type=boolean, default=0
     Set this keyword to replace all the commands in a current cgWindow or to
     create a new cgWindow for displaying this command. If the DISPLAY keyword is
     also set, a new cgWindow will be created.
xrange: in, optional, type=fltarr(2)
     A two element array giving the X range of the image. By default set to
     [0, size of image in X].
xtitle: in, optional, type=string, default=""
     The X title of the image plot. Used only if `AXES` is set.
xvector: in, optional
     A vector of X values that can be used as an alternative mthod of specifying the `XRange` of the plot.
yrange: in, optional, type=fltarr(2)
     A two element array giving the Y range of the image. By default set to
     [0, size of image in Y].
ytitle: in, optional, type=string, default=""
     The Y title of the image plot. Used only if `AXES` is set.
yvector: in, optional
     A vector of Y values that can be used as an alternative mthod of specifying the `YRange` of the plot.
_ref_extra: in, optional, type=varies
     Any keywords defined for the TV command can be used. This applies only
     if the TV keyword is set.
Returns
Returns an image that can be displaye properly.