Image Processing

Learn Photo Editing

Learn Photo Editing

Get Instant Access

Using a laptop computer in my observatory means that for processing I have to get all the files, i.e., images and calibration frames, downloaded into my desktop computer. For this I use an ethernet network, which transfers the files in just a few minutes. When they have been backed up to CDROM, they are deleted from the laptop.

For many years I used QMIPS32 for calibrating and processing images but I have switched to a combination of the freeware IRIS and MAXIM. The latter software is not cheap but makes calibration quick and easy, especially for my methodology of taking tens of exposures of each single object. If several objects were taken during an evening, it can create master bias and dark and flat frames, further speeding up processing. Each individual frame is calibrated separately before being "aligned." I do the latter using the single-star option, i.e., every frame is loaded in turn and simply by pointing at the same star on each, MAXIM

Figure 9.7. A rather extreme case of noise reduction. The Flame Nebula, taken with no-light-pollution-blocking filter on a C8 telescope (top), rescued with a wavelet filter in IRIS Software (bottom).

works out the offset and then adds all the frames together. The dynamic range that results considerably exceeds the traditional 64,000 limit, but MAXIM can handle this by saving files in 32-bit FIT format. Before saving the summed frame I remove the approximate background level, erring a bit on the low side so I don't inadvertently remove faint detail. I can then save the image as the calibrated frame.

However, actual image processing cannot yet begin. Because of the severe light pollution there will inevitably be a background gradient across the image. This varies according to where my telescope was pointing and which neighbor had his outdoor lights on! MAXIM has "flatten background" and "remove gradient" commands but I have found these unable to cope with the complexity of my light-polluted images. The only program able to rescue the image from the fog that I have found is IRIS. This has a command to synthesize a background manually (it has an auto command but manual is usually better). To do this you simply "point" at the background, carefully avoiding stars and the object of interest. Two hundred or three hundred points are usually enough. A polynomial best-fit back-

Figure 9.8.

NGC2403. Color image using the LRGB method with just a red and blue image. The green was synthesized by combining the R and B and the R used as the L. Top left - red image; top right - blue image; bottom-final LRGB image.

Figure 9.8.

NGC2403. Color image using the LRGB method with just a red and blue image. The green was synthesized by combining the R and B and the R used as the L. Top left - red image; top right - blue image; bottom-final LRGB image.

ground can be synthesized and saved. This is then simply subtracted from the calibrated image.

At last, image processing can begin. I usually try two options and see which works best. The first is Digital Development (DD) using MAXIM. This tends to be good with faint tenuous objects such as the outer reaches of galaxies, but it also holds detail right to a galaxy's core. If not carefully carried out it can produce dark rings around bright stars. No further processing is usually required with this option. The second option is Richardson-Lucy (RL) deconvolution using IRIS. This sharpens without producing dark rings round stars. It is an iterative process and restores the image based on a selected star within it. I generally only use 5 to

Figure 9.9. Flaming Star Nebula IC405. This color image is a combination of a full-resolution CCD mono image and low-resolution color information from an old 35mm slide.

Figure 9.9. Flaming Star Nebula IC405. This color image is a combination of a full-resolution CCD mono image and low-resolution color information from an old 35mm slide.

Figure 9.10. IC342/Maffei Galaxy Group montage (1). This group has some of the most obscured galaxies known and it is only relatively recently that several members have been discovered. Yet our backyard scopes, under light-polluted skies, can successfully image them in the infrared.
Figure 9.11. IC342/Maffei Galaxy Group montage (2). Red/infrared images with exposure totaling 45 minutes (90 x 30 seconds).

10 iterations - any more and noise tends to become objectionable. It then requires a "stretch" to make the faint detail visible. Logarithm is usually too strong and a gamma stretch with a value of 0.5 is generally about right. If I cannot decide which option is better (a frequent dilemma), then I sit on the fence and take an average of the DD and RL results!

Light pollution manifests itself by producing images that are grainy or speckled. This is the random element that cannot be removed. Now for that smarter processing - the magical noise-reduction filter. By filter I mean a digital process, not a special piece of glass. For this I have found IRIS's commands the best (Adaptive and Wavelet). These work well on diffuse nebula and elliptical galaxies where lack of smoothness in the image is all too obvious. They turn our grainy light-polluted images into those (almost) matching the smoothest ones from the darkest of sites (see Figure 9.7).

Figure 9.12.

Selection of Palomar Globular clusters. Pal 3 and 4 are extremely remote.

Figure 9.12.

Selection of Palomar Globular clusters. Pal 3 and 4 are extremely remote.

Was this article helpful?

0 0
Photoshop CS Mastery

Photoshop CS Mastery

Artists, photographers, graphic artists and designers. In fact anyone needing a top-notch solution for picture management and editing. Set Your Photographic Creativity Free. Master Adobe Photoshop Once and For All - Create Flawless, Dramatic Images Using The Tools The Professionals Choose. Get My Video Tutorials and Retain More Information About Adobe Photoshop.

Get My Free Videos


Post a comment