Christoph PetermannHomepage

 Astro Image Processing

Last Update: 1 June 2001

This page contains my worksheet for my Astronomical imageprocessing. Since consequently following these schemes I get good results. This work does take a lot of time. I also optimized older images made with my webcam with very good results. The intermediate space used on a harddisk for producing on single image easily reaches several hundred megabytes.

I. Preparations


I usually generate the following file hierarchy:

astroimages/[date]/ast/ containing raw .ast CCD camera files like


make there a working directory d:/astroimages/2001-03-12/work where I make all processes.

make a result directory like: d:/astroimages/2001-03-12/result

I try to keep this format every time.

CCDCONV OES .AST format to Fits .FTS. The Program is by Andre Wulff. As the program does not work correctly you have to flip or mirror the image at the end of the process.


rename manually "0000000x.fts" to "000000x.fts" for 01 ..09, because IRIS needs:
and OES generates a leeding "0"

II. Processing planetary images

with the use of IRIS

Find the 0 = Black level offset
Add a constant (negative) to the image sequence= 0 leveloffset2 000000 a [offset] [num]
Find the maximum brightness level (MaxBright = ca 4000)mult2 a b [coeff] [num]
Multiply the sequence with 32000/MaxBright
Find and sort the best imagesbestof b
Select and sort outselect b c
register planetary (2)
alt : menu !
pregister c d [size=256] [num]
Number I want to add i.e. NA = 10 (example)
Determine [coeff2] = 1/NA
Multiply the sequence with 32000/MaxBrightmult2 d e [coeff2] [num]
Add all imagesadd2 e [NA]
Save the result firstsave opt_raw(.fts is assumed)
option multiply the sizescale 2 x y
multiply the size (for OES LcCCD07)
to get a square image
scale 2 1.33 1.551
Optimize the image
repeat ... until the image looks goodunsharp 1.5 12 (down to 0.7 1.5)
gauss 0.8 .. 1.0 for noise reduction
flip the image horizontal as CCDCONV does not convert correctly

III. Deep Sky image processing

Deep Sky processing I do very similar to the planetary processing. Even it seems to be a lot of work, I consequently follow these steps.

M27This is a raw image from the camera with no darkfield subtracted.
M27This image I have processed with PictureWindow 2.0 in May 2000
M27 After I got familiar with IRIS, I came to this result showing much more details and fainter stars than the earlier result above.
Get the first image and repeatload 00000001.fts
For my small images I have to resample and adjust the images to get a square 256*256 image.
As this does not work for a sequence, I do it for each image manually.
scale 2 1.333 1.551
save image in a sequence using a [generic name]save a1 ; next time save a2
Find the 0 = Black level offset (if necessary). I subtract the dark field in the CCD camera program already.
alt : menu !
register a b [size=256] [num]
Number I want to add i.e. NA = 10 (example)
Add all imagesadd2 b [NA]
Save the result firstsave opt_raw(.fts is assumed)
Optimize the image
repeat ... until the image looks good0.8 2.0
gauss 1.0
flip the image horizontal as CCDCONV does not convert correctly

IV. The tools I use

There are quite a number of tools available on the market, but I use only a few, because it is not necessary to use many different programs. After working with Picture Window 2.0 ( for a while, I came to IRIS by C.Buil. This software is free and does THE job. But many other programs do also a fine job. Stay with that one you like best.

 Text and All Images are Copyright by Christoph Petermann DF9CY

GO (back) and visit my homepage