Newsgroups: comp.compression
Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!magnus.acs.ohio-state.edu!news
From: henden@hpuxa.acs.ohio-state.edu (Arne A. Henden)
Subject: Re: Atronomical data compression
Message-ID: <1991Mar27.021241.6339@magnus.acs.ohio-state.edu>
Keywords: Spectra, Keck
Sender: news@magnus.acs.ohio-state.edu
Nntp-Posting-Host: hpuxa.acs.ohio-state.edu
Organization: The Ohio State University
References: <1991Mar23.013557.28151@nntp-server.caltech.edu>
Date: Wed, 27 Mar 1991 02:12:41 GMT
Lines: 33

In article <1991Mar23.013557.28151@nntp-server.caltech.edu>
 sns@deimos.caltech.edu (Sam Southard Jr.) writes:
>I have a topic/question that should be suitable for this newsgroup.
   [ ..much deleted!]
>[compressing astronomical images]
> Obviously, if this kind of data is going over the
>Ethernet, we want to compress it as much as possible.
>
>Does anyone have any suggestiongs?
>
>Sam Southard, Jr.
>{sns@deimos.caltech.edu|{backbone}!cit-vax!deimos!sns}

  Most astronomical images rattle around some low value with
a few spikes reaching near saturation (stars).  Others have
proposed some interesting techniques, such as using a difference
model.
  One technique that we wanted to try, but have never taken
the time to program, is to use bit plane compression.  For the
example above, divide the image into 16 1-bit images and
then run-length compress each image.  You will find a very
high compression ratio on the high order bits and essentially
zero compression on the low order bits.  I'd bet the average
would be about 40 percent compression, but the task would
be quite CPU intensive.
  We were originally going to use this technique for our
remote observing quick-look images, but find that run-length
compression with an adjustable threshold works well enough
for now, and intend to buy a pair of boards using the c-cube
chip in the near future to try a different approach.
Arne Henden  Astronomy Dept.