Using Contrib Color Look-Up Tables
CIAO 4.17 Science Threads
Overview
Synopsis:
Users can follow this thread if they wish to use the contributed color look-up tables in ds9, matplotlib, or dmimg2jpg.
Purpose:
One of the easiest customizations that users can do in CIAO is to use a non-standard color look-up table (LUT) or some times called color map.
Beyond the aesthetics of a particular color choice, different LUT's can be used to convey additional meaning: highlighting certain regions, emphasizing gradients, and provide additional reference between datasets.
This thread will show you how to load the contributed color look-up tables into ds9, matplotlib, and dmimg2jpg
Related Links:
- An inf@Vis article on the use of colour in cartography and visualisation, as championed by by Cynthia A. Brewer of Penn State, who created the ColorBrewer site.
- A follow-up article in inf@Vis on the spectral (or rainbow) palette, pointing out that the Rainbow Color Map (Still) Considered Harmful.
- mean_energy_map tool
- dmimg2jpg tool
- ds9 thread
Last Update: 13 Jan 2022 - Review for CIAO 4.14. No changes.
Contents
- Get Started
- Background Information
- Contributed Look-Up Tables
- How to Use Contributed LUT's
- History
- Images
Get Started
Download the sample data: 214 (CAS A)
unix% download_chandra_obsid 214 evt2
Background Information
Many different fields of science rely on different color look-up
tables (LUT's): medical
imaging, cartography, meteorology, fluid dynamics,
as well as astronomy. It is not surprising then that
there are many different color maps
available. The largest repository seems to be
cpt-city
which has 100's of different color maps from both the scientific
community as well as the realm of artistic/graphic-design. ds9
version 7 now has a direct link to cpt-city and all their color
maps are now available in ds9's .sao
format.
The color map chosen is up to the user to decide based on what information they are trying to convey. For most scientific purposes, the only common restriction would be that colors are unique: having the color "white" show up at both a low intensity level and a high pixel intensity level in the color map may lead to ambiguity. This is not an absolute rule as one may choose to obscure certain regimes of the data with the same color to draw focus on a different area.
Examples of which color map to use when
If one were to have an image whose pixels represent the incident integrated flux, a colored gradient, one maybe chosen to symbolically match the energy band pass, may be a good choice (Figure 1 and Figure 2). An image showing temperatures is often shown using a rainbow color scheme (Figure 3). And an image showing color, that is the ratio of two images in different energy bands, may best be shown with a color map that smoothly varies between two colors (Figure 4).
[Version: full-size]
Figure 1: Image of Low Energy Events
[Version: full-size]
Figure 2: Image of High Energy Events
[Version: full-size]
Figure 3: Pseudo Temperature Map
[Version: full-size]
Figure 4: Hardness Ratio
Contributed Look-Up Tables
These LUT's were extracted from two different software package and the data were reformatted to ds9's .lut format.
- HEAsoft's
XImage
package,
a multi-mission X-ray image display and analysis program
- National Institutes of Health's
ImageJ
package,
a public domain Java image processing program inspired by [National Institutes for Health] Image
.
Note: Some of these color maps may be identical or nearly identical to one or more of those already available in ds9, matplotlib, or dmimg2jpg
XImage
Click on colorbars to see example. |
ximage_bb blue1 blue2 blue3 blue4 bluebase1 bluebase2 bluebase3 bluebase4 bluebase5 bluebase6 bluebase7 bluebase8 bluebase9 bluyel brown eyellow1 gray1 gray2 green1 green2 green3 |
green4 green5 green6 green7 green8 green9 invgray invgray1 invspec pink1 pink2 pink3 pink4 pinkbase1 pinkbase2 pinkbase3 pinkbase4 pinkbase5 purby purple1 purple2 purple3 |
purple4 puryb red1 red2 redbase1 redbase2 redbase3 redbase4 redbase5 redbase6 rgb4 rgb6 spectrum vt1 vt2 vt3 vt4 vt5 vt6 vt7 yellow1 yy |
ImageJ
Click on colorbars to see example. |
000-gray 001-fire 002-spectrum 003-ice 004-phase 005-random 16_colors 16_equal 16_ramps 20_colors 32_colors 5_ramps 6_reserved_colors 6_shades amber auxctq blue_orange blue_orange_icb brain brgbcmyw cells cequal cmy-cyan |
cmy-magneta cmy-yellow cmy cold cti_ras edges gem-16 gem-256 gold gyr_centre heart hue hue_ramps_08 hue_ramps_16 icool iman invert_gray isocontour log_down log_up mixed neon-blue neon-green |
neon-magenta neon-red pastel rgb-blue rgb-green rgb-red royal sepia siemens smart split_blackblue_redwhite split_blackwhite_ge split_blackwhite_warmmetal split_bluered_warmmetal system_lut thal_16 thal_256 thallium topography unionjack vivid warhol |
How to Use Contributed LUT's
All the color look-up tables will be located in $ASCDS_CONTRIB/data. You can access them by name by using the ximage_lut.par and imagej_lut.par parameter files.
unix% plist ximage_lut Parameters for /soft/ciao-4.4/contrib/param/ximage_lut.par # # Color Lookup Tables extracted from NASA XImage package # # http://heasarc.gsfc.nasa.gov/xanadu/ximage/ximage.html # # converted to the ".lut" format used by ds9, chips, and dmimg2jpg # # (bb = ${ASCDS_CONTRIB}/data/bb.lut -> /soft/ciao-4.4/contrib/data/bb.lut) Color Lookup Table (blue1 = ${ASCDS_CONTRIB}/data/blue1.lut -> /soft/ciao-4.4/contrib/data/blue1.lut) Color Lookup Table (blue2 = ${ASCDS_CONTRIB}/data/blue2.lut -> /soft/ciao-4.4/contrib/data/blue2.lut) Color Lookup Table (blue3 = ${ASCDS_CONTRIB}/data/blue3.lut -> /soft/ciao-4.4/contrib/data/blue3.lut) Color Lookup Table (blue4 = ${ASCDS_CONTRIB}/data/blue4.lut -> /soft/ciao-4.4/contrib/data/blue4.lut) Color Lookup Table ... unix% plist imagej_lut Parameters for /export/ciao-4.4/contrib/param/imagej_lut.par # # # Color Lookup Tables extracted from the NIH's ImageJ package # # http://rsbweb.nih.gov/ij/ # # converted to the ".lut" format used by ds9, chips, and dmimg2jpg # (000-gray = ${ASCDS_CONTRIB}/data/000-gray.lut -> /export/ciao-4.4/contrib/data/000-gray.lut) Color Lookup Table (001-fire = ${ASCDS_CONTRIB}/data/001-fire.lut -> /export/ciao-4.4/contrib/data/001-fire.lut) Color Lookup Table (002-spectrum = ${ASCDS_CONTRIB}/data/002-spectrum.lut -> /export/ciao-4.4/contrib/data/002-spectrum.lut) Color Lookup Table (003-ice = ${ASCDS_CONTRIB}/data/003-ice.lut -> /export/ciao-4.4/contrib/data/003-ice.lut) Color Lookup Table ...
Using contributed LUT's in ds9
As with most things in ds9, there are at least 3 different ways to load a new color map: via the GUI, on the command line, and via XPA. Only the first two are shown here.
On the command line, you use the -cmap load command
unix% ds9 acisf00214N000_evt2.fits -cmap load $ASCDS_CONTRIB/data/heart.lut
[Version: full-size]
Figure 5: ImageJ "heart" LUT in ds9
Alternatively, in the ds9 GUI, you can load the color map by going to Color -> Colomap Parameters ... In the sub-window, then go to File -> Load Colormap ..., navigate to the CIAO contrib/data directory and select the file name.
When a custom color map is loaded, it will be available under the top level "Color" menu. Prior to version 7, the name of the new color map would simply be appended to the list of those built-in. In version 7, custom color maps have their own sub-menu.
Using contributed LUT's in matplotlib
In this example we use the parameter module to retrieve the file name from the imagej_lut.par file. We then parse this file into the red, green, blue, and alpha (transparency) channels used by the ListedColormap class.
% python import paramio as pio from pycrates import read_file import matplotlib.pylab as plt from matplotlib.colors import ListedColormap heart_lut = pio.pget("imagej_lut", "heart") tab = open(heart_lut,"r").readlines() rgb = [ x.strip().split() for x in tab if not x.startswith("#")] rgba= [ [float(x[0]),float(x[1]),float(x[2]),1.0] for x in rgb ] heart_lut_cmap = ListedColormap(rgba) img = read_file('img_sm_asinh.fits') imgvals = img.get_image().values plt.imshow(imgvals, origin='lower', cmap=heart_lut_cmap) plt.savefig("matplotlib.png")
[Version: full-size]
Figure 6: ImageJ "heart" LUT in matplotlib
Using contributed LUT's in dmimg2jpg
The look-up table can be fed directly to dmimg2jpg. In the follow example we make use of a parameter redirect to locate the file name.
unix% dmimg2jpg infile=img_sm.fits outfile=img_sm.jpg \ lutfile=")imagej_lut.heart" \ scalefun=log mode=h clobber=yes
[Version: full-size]
Figure 7: ImageJ "heart" LUT in dmimg2jpg
History
10 Oct 2012 | Initial version. |
13 Dec 2012 | Review for CIAO 4.5; added note about adding color bar via chips gui and added several inf@Vis articles on the use of color to the Related Links section. |
24 Apr 2013 | To avoid a naming conflict, ximage.par and imagej.par have been renamed ximage_lut.par and imagej_lut.par |
25 Nov 2013 | Review for CIAO 4.6. |
16 Dec 2014 | Reviewed for CIAO 4.7; no changes. |
01 Feb 2016 | Updated ds9 links. |
29 Dec 2016 | Reviewed for CIAO4.9. No changes. |
28 Mar 2019 | Updated to use matplotlib. |
13 Jan 2022 | Review for CIAO 4.14. No changes. |