Astronomical images are often displayed in false color, i.e. intensity levels in the image are mapped onto a table of colors that are easily and unambiguously ordered by the typical viewer's mind. Such color coding can be highly effective in conveying subtle differences in flux because the human eye is much more sensitive to differences in hue than it is to differences in brightness.
Sometimes however one has more information about a scene than just a simple flux map. Perhaps one has multiple flux maps taken at different wavebands. Or, perhaps one has data that specifies some second property (e.g. median energy, or hardness ratio) at each point in the flux map. In these cases one often wants to try to construct a ``true color'' image which encodes both the flux and ``extra'' information at the same time. The true_color_image tool, provides such capabilities.
The most common situation is to have three co-aligned images from different bands that you wish to represent with the red, green, & blue components of a color image. If the images are in IDL variables then call the true_color_image tool directly (Section 8). If the images are stored in FITS files then use the fits_viewer tool (no parameters) to load the three images into a dataset_2d tool, then select Analysis 
Combine Images to spawn a true_color_image tool.
Other tools that construct 3-color images commonly allow/force adjustment of the brightness scaling ranges of the 3 components components independently. This leads to unnecessary distortions in the ``hue'' (color) in the resulting image. The most noticable example of this is that pixels with a large signal in all three bands are rendered as white, completely discarding any information about the actual ratios among the three components.
The true_color_image tool avoids these distortions via a brightness scaling approach that preserves the ratios among the three components (the ``hue''). For example, a pixel whose red component is significantly larger than the green & blue components will appear red in color, no matter how the brightness of the image is scaled. The tool applies a user-adjustable gain to the three component images to accommodate component images that intrinsically have different signal strengths. Adjusting these gains controls the overall color balance of the image.
The true_color_image tool provides both a ``luminous'' color model where zero signal is rendered as black, and a ``subtractive'' color model model where zero signal is rendered as white. To illustrate, consider the three (red, green, blue) component images shown in Figure 11. These data contain 512 unique combinations of the three components, i.e. 512 different ``hues''. The corresponding true color images are shown in Figure 12.
 |
A true color image may alternatively be derived from two components - one component directly controls the hue of the result and the second component directly controls the brightness. Typically the brightness component would be a standard intensity image (# of counts detected per pixel) and the hue component would be a statistic map (see Section 6), for example a median energy map.
Keep in mind that statistic maps can be corrupted by high background, as described in Section 6.
Please note that the data in the True Color Image tool is NOT automatically updated when the constituent images in other widgets are changed.
