K3's Astronomy - Deep Sky
"When I consider your heavens, the work of your fingers, the moon and the stars which You have set in place, what is Man that You are mindful of him?" -- Psalm 8:3,4

M57 - The Ring Nebula - New experiences with getting a photograph
Date, time: 2.10.2002 (20:00 - 21:36 UT)
Resolution: 0.96 arcsec/pixel
Camera: Vesta 675 SC2
Telescope: 8" F6 Orion Europa Newtonian
Result focal ratio: F6
Exposure: 129x30s (gain 100%, Amp OFF, outdoor white balance), 16xDF
Captured and processed by K3CCDTools, then in Maxim DL Demo and IRIS.
Seeing: very soft haze, lim.magnitude for naked eye 4.2, temperature 10C

There are several key factors, which helped me to achieve better photos during 1 year:.

A. Enhanced equipment

In comparison to my first imaging of M57, I used enhanced equipment:

1, Peltier cooled camera

For imaging I used Vesta 675 SC2 webcam modified according to Steve Chambers with Peltier cooling system.
In 30s exposure, 100% gain and ambient temperature 10C there were virtually no hot pixels:

You can compare it to dark frame of 5s, 75% gain exposure with not cooled system.

2, Homemade autoguiding system

Long exposures require good and stable mount. As my GEM1 mount is only relatively cheap entry level mount, it has a big periodic error. I build my own auto guiding system, which guarantees me periodic error maximum +/-5arcsec (standard deviation 1.6arcsec).
The efficiency of autoguiding shows the following graph:

B. Frames aligning

Better aligning algorithm in K3CCDTools allows to use stronger histogram stretching. When histogram is heavily stretched, then also small artifacts in image are visible. Well known is problem of strips:

This is a summed image of 129 frames aligned by using FFT correlation method.

This is a summed image of 129 frames aligned by using automatic FFT correlation and then fine aligned by Difference Alignment. What a difference!
The trick is using a small Radius (2 pixels) and Threshold higher than background level. In that case soft strip pattern (usually not visible by eye in raw frames) in background has no effect for aligning and frames are aligned only according to bright objects (stars).

It is interestig, that analysis of difference between these 2 mentioned methods showed, that alignment differents were not higher than 1 pixel!

C. Post-processing results

Processing details:
1, aligned and summed in K3CCDTools
2, histogram stretching in K3CCDTools
3, unsharp mask 1, 0, 50

The result image was produced only by using K3CCDTools without need of any other software package.

Processing details:
1, aligned and summed in K3CCDTools, then exported to FIT
2, histogram stretching in Maxim DL Demo
3, cutted red channel's low end of histogram (light pollution) in Corel PhotoPaint
4, adaptive unsharp mask (66%) in Corel PhotoPaint

Stars are not perfectly rounded, because of not perfect seeing and not perfect mount. Unsharp masking helped to enhance the image, but stronger unsharp mask causes higher noise in image.

I started to think of using some image deconvolution.
This image was processed by Maximum Entropy Wizard in Maxim DL Demo.
Processing details (the input was the above image):
1, Maximum Entropy for each R-G-B channel in Maxim DL Demo
2, Unsharp mask 2,0,100 in Corel PhotoPaint
3, Unsharp mask 1,0,100 in Corel PhotoPaint

Maximum Entropy Wizard enhanced some details in nebula, but two 15.0 mag stars above nebula are not as distinguishable like in above image.

Processing details:
1, aligned and summed in K3CCDTools, then exported to FIT
2, histogram stretching in Maxim DL Demo
3, low end histogram cut for R channel in IRIS
4, Richardson-Lucy deconvolution for each R-G-B channel in IRIS
5, combining R-G-B plates into TrueColor bitmap in IRIS

This is my final version of processing. Richardson-Lucy deconvolution algorithm in IRIS did a great job. It enhanced nice details in nebula area.

This is a special version of above image for laptops screens - they tend to enhance dim parts (LCD screens seems to have higher gamma than standard cathode ray tube monitors).
The low end of histogram was cut.

Negative of IRIS version - better for finding dim objects.

This is a special processing (histogram stretching) for enhancing dark details.
IC1296 is a spiral galaxy.

Back to M57

Computer generated images, real images, drawings and texts are property of the author and may not be reproduced or used without permission of author.


Last Update: 3.10.2002