Stretching levels and colors
The computation of the logarithm of an image is a classical method for compressing the dynamic range. The operation emphase simultenaously information in faint and bright levels of the image.
From the View menu, run the Logarithm command and adjust the visualization threshold:
Part of nebula NGC 7635 (FSQ106
+ moded EOS350D). Left, linear distribution of pixel intensities. Right,
For an identical effet from the command line, enter:
The argument is the max value in the final image.
After histogram equalization, the intensity distribution is uniform. Modified equalization is an improved version of standard equalization for better result. For example the command Modified equalization of View menu can be used for a better view of darkness part of the "observatory image":
Up, the original image. Down, the same image after modified equalization.
Dynamic stretching command (nearly similar to the well know Digital Processing Development) of View menu is an interactive method for modifiy transcoding of pixel intensities. The function is optimized for deep-sky images.
The constant value is a pedestal added for skip overflow calculation in some situations. If abnormal pixels values appear, try constant 100, 200, 300...
Open the dialog Color stretching of the View menu, then
The same effect from a console command (ASINH):
>ASINH 0.0102 1.43
ASINH command stretch the image for show faint objects, while simultaneously, preserve the structure of bright objects of the field. The color contrast is boosted by the application a non-linear stretch: the Arc Sinus Hyperbolic function. This method permit to reveal an enormous amount of information: index color of stars, faint nebulae, galaxies having a distinctive colors (see for example the Hubble Deep-Space images of the HST, many time processed with a function very similar to ASINH). The arcsinh is a new manner of defining the magnitudes scale, see R. Lupton, Astronomical Journal, 118, 1406-1410. This scale magnitudes has properties very interesting when one applies to color images because it boost the colors index of the objects (see also R. Lupton & all, PASP, 116,133-137). The colors contrast is very strongly accentuated whereas the noise increase is contained.
The parameter [ stretch] permit to adjust the non-linearity factor. A null value corresponds to a standard linear scale. The characteristic values go from 0.001 to 0.1. The parameter [intensity] adjust the intensity of the final image. The characteristic value for this parameter goes from 1 to 50 (carry out tests and exploit the visualization thresholds).
Here some application examples.
Look the color image of galaxy NGC 1055. It was produced starting from three distinct CCD frames taken through filters red, green and blue (photometric RVB system) with the 60-cm telescope of Pic du Midi. Visualization is traditional, linear in intensities. The nuclear region appears saturated.
Below the same object, but after having used command ASINH 0.01 20. (or the equivalent slide position in the dialog box Colors Stretching). The core is less saturated and the extensions of the galaxies become visible:
Now, the result of command ASINH 0.05 20. At present, at the same time the core and the extensions are visible simultaneously whereas the noise is only moderately accentuated. Note the trail of satellite, presents in the red image, not easily perceptible with a linear scale, but clearly visible in the image processed with the command ASINH (very significant boost of color contrast).
The example below shows an image RVB composition taken with a Takahashi Epsilon 160 (M27). On the left, linear visualization, on the right, arscinh visualization.
Successively a composite image of the cluster M37 with linear visualization and after ASINH command (Audine CCD camera and refractor Takahashi FS-106). Now, stars show a variety of red and bleus colors. Two star populations are clearly revealed.
The last example shows the application of the arcsinh technique applied to M17 (60-cm telescope of the Pic du Midi). From top to bottom, linear display, logarithmic display (command Logarithm of View menu), and arcsinh stetching.
Applied an opacity maskThe opacity masking is a method used for makes image easier to view on a video monitor. It is for example a tool for displaying the whole dynamic of an image, which is a current situation in astronomy. You can thus observe the details of the nuclear part of a galaxy while preserving the aspect of the spiral arms (see exemple below). But, it is very important to note that this filter is strongly nonlinear, which must be used with prudence, only for the matter of some visualization, because many artifacts are gererated if it is employed with force. Be careful!
The operation proceeds in two phases:
(1) You must first create the opacity mask. The more intense is, the more it will obscure the corresponding parts of the image to be processed. Generally, the mask is made starting from the image to even treat it, by producing a gaussian blur.
(2) The mask is applied to the image. The effect can be adjusted interactively.
We apply the opacity mask to the below image for emphasize information of the darkness zones...
Generate the mask by bluring the image. Use the gaussian blur for this:
Save the mask on the disk (give the name "mask" for example), reload the image to process and applied the mask. For this operation, run the command Opacity mask of View menu:
Enter the mask name and adjust the slider. The result is
Below, the application of the filter for an image of galaxy:
Command SBLUR command generates a blur in the image with a force more higher since the objects are more intense. This function is used on star fields to produce an Akira Fujii effect, for example to reveal the contour of the constellations.
The syntax is : SBLUR [SIGMA] [GAMMA]
The parameter SIGMA adjust the degree of the blur (select values between 2 and 15) and the parameter GAMMA makes adjust the brightness of intense stars (the characteristic value is between 4 and 10). For an optimal effect it should be take care that the most brilliant stars are not saturated. This command boost the colors and can be also used in combination with "asinh" method.
For example, apply SBLUR command to a wide-field Sagittarius-Scorpius captured with a Canon EOS 10D and a 17 mm objective lens at f/4 (composite of 8 images exposed 5 minutes at 400 ASA), from the observatory of the Pic du Midi observatory (July 2004). Here the image with a standard display:
The same image after application of command:
>SBLUR 5 7
It is noticed that the weak stars keep their sharpness whereas the brilliant stars are well raised. The Scorpius constellation is now for example easily identifiable. By modifying the parameters, by successive tests, it is easy to change the effect.
Another example, the Cassiopeia region before processing (Canon EOS10D + 17 mm lens image):
The Cassiopeia region after selective blur:
Open the dialog Saturation adjustment of the View menu and act on the cursor. Here for example, on the left, a standard image of the Moon carried out with a Canon EOS 5D associated to a Canon 400 f5.6 lens and a x2 extender. The image is a stack of 6 frames. On right-hand side, the saturation was very strongly enhanced by using the tool twice successively with the cursor on position 4 each time. The true colors of lunar surface are thus revealed (it is a geological map of Lunar surface).
The saturation tool.