DSP Based Telescope?

[mrzee]

Here's possibly a dumb idea but I wouln't mind a few opinions on the problems I'd be facing if I attempted to try and build one.

I've always wanted a teleccope, however:

1) A decent one is $$$$
2) I wanted to look at different parts of the spectrum not just the visible.

With current technology comes the possibility of either new ideas or just made afordable ideas.
Using a CCD element in conjunction with a reflector/lens it is possible to construct a crude telescope cheaply. By crude I refer to the precision of the reflector and the lens. Using DSP software to correct the inaccuracies can allow the posibility of producing good quality images from the rather cheap physical components. At the very least cheaper.

The question is how crude could I go? Can I use a fiberglass/plastic reflector from say a satelite receiver after applying a reflective coating? Or would I need to build something specificaly for the purpose. If so how precise would it need to be? My guess is the worst case error can only move +/- 50% of the CCD visible area over the whole reflector area for parrallel beams of light. This would be pretty crude! If the reflector was mapped it could still be reconstructed using DSP (be it very slow in this case).

Why do all this?
1) it's fun, challenging, interesting
2) It may produce cheap but good results for home constructors to build.
3) It offers the IR spectrum and/or radio waves as standard.

Any thoughts welcome.

[julesstoop]

How would you map the reflector to an accuracy of about half a micron, and subsequently keep it to the measured state?

[t00fri]

Here's possibly a dumb idea but I wouln't mind a few opinions on the problems I'd be facing if I attempted to try and build one.

I've always wanted a teleccope, however:

1) A decent one is $$$$
2) I wanted to look at different parts of the spectrum not just the visible.

With current technology comes the possibility of either new ideas or just made afordable ideas.
Using a CCD element in conjunction with a reflector/lens it is possible to construct a crude telescope cheaply. By crude I refer to the precision of the reflector and the lens. Using DSP software to correct the inaccuracies can allow the posibility of producing good quality images from the rather cheap physical components. At the very least cheaper.

The question is how crude could I go? Can I use a fiberglass/plastic reflector from say a satelite receiver after applying a reflective coating? Or would I need to build something specificaly for the purpose. If so how precise would it need to be? My guess is the worst case error can only move +/- 50% of the CCD visible area over the whole reflector area for parrallel beams of light. This would be pretty crude! If the reflector was mapped it could still be reconstructed using DSP (be it very slow in this case).

Why do all this?
1) it's fun, challenging, interesting
2) It may produce cheap but good results for home constructors to build.
3) It offers the IR spectrum and/or radio waves as standard.

Any thoughts welcome.

Times seem to be changing;-). I am involved with amateur astronomy since the age of 12 when I built my first single lens refractor (5cm;-)). Before I was 20, I had ground several 20cm = 8 inch mirrors myself in the cellar. All it costed was the price of the raw glass and the aluminizing, together with patience of course. Doing an optical precision instrument to an accuracy of about 1/10 of the wavelength of yellow light (!) with a little more than "kitchen utensils" is a great adventure by itself and checking out that precision involves most interesting physics!

It seems that the option of making things yourself is getting lost and people constantly argue they don't have the time for it;-)...

Bye Fridger

[selden]

But Fridger, this is a "do-it-yourself" project! It's just that it's primarily electronics and/or software instead of big, heavy chunks of glass and grit and tar and water and... Sky & Telescope used to publish a 3 book series on "Amateur Telescope Making". I found them to be fascinating reading. But all that physical labor!

Something like this was done in order to clean up the original images from the HST before the optics were fixed. I know next to nothing about the algorithms used except that they depended a lot on knowing in advance what kind of object was being observed. It's relatively easy to clean up a starfield that's only supposed to contain circular spots. Cleaning up an image of the "random" lightfield from a nebula is somewhat more difficult.

[t00fri]

But Fridger, this is a "do-it-yourself" project! It's just that it's primarily electronics and/or software instead of big, heavy chunks of glass and grit and tar and water and... Sky & Telescope used to publish a 3 book series on "Amateur Telescope Making". I found them to be fascinating reading. But all that physical labor!

Something like this was done in order to clean up the original images from the HST before the optics were fixed. I know next to nothing about the algorithms used except that they depended a lot on knowing in advance what kind of object was being observed. It's relatively easy to clean up a starfield that's only supposed to contain circular spots. Cleaning up an image of the "random" lightfield from a nebula is somewhat more difficult.

Selden,

I did not react to the original proposal, since I did not see a reasonable chance for it to work satisfactorily. For that reason I also went back to the classical "Amateur Telescope Making" philosphy, since that definitely works...(I own those three ATM volumes of course and many more...).

There are 2 things that seem "bad":

--the field subtended by an affordable CCD in the focus of whatever focussing element is tiny

--I cannot see how one can hope to correct for miserable imaging due to a plastic mirror by means of software. This strategy I expect only to work so long as the corrections remain modest as compared to the basic imaging quality! Nature/physics cannot be cheated;-))

Sorry, that plan is beyond my capabilities;-)

As to the HST story, this was different: they had essentially forgotten about the absence of gravity under actual operation conditions. The HST mirror was figured on earth under the influence of gravity. This meant a deformation of the mirror's curve in space, since gravity was not pulling anymore;-)

These effects are devastating on the image quality, but in size minute as compared to the trash imaging from a plastic mirror;-)

Bye Fridger

[selden]

Fridger,

I don't think I was disagreeing with you, just making comments

Somewhere on the stsci.edu site are (or were) pictures comparing the three states of the HST: before software corrections, after software corrections and after optics corrections. (Hopefully they'll return after the Mars excitement has toned down.) They show quite well the things one can aspire to doing.

[added slightly later]
The Hubble optics problem was not so subtle as you mention: the mirror was given an amazingly perfect figure, but with gross spherical aberration due to a problem in Perkin-Elmer's null testing hardware.

A popularized description of the mistake is at http://news.bbc.co.uk/2/hi/science/nature/638187.stm

A more formal, detailed description is at http://history.nasa.gov/SP-4219/Chapter16.html

I have not yet found a copy of the report from the investigation. Apparently it was a result of management problems which were compared to those causing the Challenger disaster.

[t00fri]

Fridger,

I don't think I was disagreeing with you, just making comments

Somewhere on the stsci.edu site are (or were) pictures comparing the three states of the HST: before software corrections, after software corrections and after optics corrections. (Hopefully they'll return after the Mars excitement has toned down.) They show quite well the things one can aspire to doing.

[added slightly later]
The Hubble optics problem was not so subtle as you mention: the mirror was given an amazingly perfect figure, but with gross spherical aberration due to a problem in Perkin-Elmer's null testing hardware.

A popularized description of the mistake is at http://news.bbc.co.uk/2/hi/science/nature/638187.stm

A more formal, detailed description is at http://history.nasa.gov/SP-4219/Chapter16.html

I have not yet found a copy of the report from the investigation. Apparently it was a result of management problems which were compared to those causing the Challenger disaster.

Selden,

much of the HST story on that URL actually matches what I have said, too. Originally it was reported that the excessive spherical aberration was due to the forgotten effect of the "missing gravity" on the mirror shape. Perhaps later, they discovered that lost spot of paint;-). This I did not know. Using Null correctors is a standard procedure for shape controlling aspheric mirrors of any kind. I have used this method, too. Nothing special. Only the sensitivity of the HST null test was of course outstanding and for reasons of effectiveness, a great amount of automatization was required. The main new point for me was the /reason/ for the fake in the null test that you located on that URL.

Bye Fridger

[mrzee]

Fridger,

W.R.T. the HST, I always thought it was that they applied the compensation twice. The specification already had the effects of gravity compensated and the manufacturer then applied it again. Just what I've always been told but hey it wouldn't be the first time I've got the wrong story.

Sometimes I don't make myself too clear and so cause confusion, so I'll try again.

W.R.T. the DSP based telescope: The examples I used were as described in the text "crude", the question was more "how crude" is it possible to go so such a technique could effectivily be used. After that was the question would the addition of a CCD and some DSP software be cost effective.

The way I tend to look at things is from the boundary conditions, in this case the best cenario is a theoretical "perfect mirror", the worst case is the cheap plastic. Adding a CCD element to even a perfect reflector has its advantages: it can see IR, imediate transfer to a computer, no camera adapters if a record is to be preserved etc. There are software packages to compensate for out of focus and perform other effects. While some may argue the merit of the effects its pretty hard to argue that image enhancement won't work as it's been done every day in hospitals CAT, MRI, X-Rays etc. The question is more how much benefit will be gained.

I totally agree with the cheap plastic option, it has more problems than solutions. To get results from a device like this it's more a question of stability of the reflector rather than the precision of the reflector, within limits of course. The main reason is that if the results are repeatable they can be filtered out with compensation. Look at refractors, even the best ones have distortions w.r.t colors as the focuing can be made only at two wave lengths. There is no reason that this can't be compensated for in software. Every telescope I know of has a lens of some sort in the light path so this can be corrected. Most systems employ multiple lens's to reduce this effect, each lens reflects a small quantity of light back so is lossy. Whilst coatinigs help all these still add up, whether significantly is a matter of opinion.

I'm not too sure on what you refer to with the focus of the CCD element. As far as I figure, a 1024x768 CCD is quite cheap now, if anyone has their screen set to this resolution and looks at an image of a planet on the display they would be impressed. If they saw something like that in their telescope they would be more than happy! N.B. the point there is the CCD element isn't the limiting factor, it's the air we breathe thats the problem. I really don't understand the reference to the focus of the CCD so I refered to the resolution. I just can't imagine not being able to focus one as it's already been done in the HST and many CCD cameras now can be attatched to telescopes. If I missed your point please elaborate as it's something that I'm not aware of then.

Having had a bit of a discussion with a friend (a physisist, dam they are great to talk to!) he correctly pointed out that the limiting factor for the presision of the reflector and CCD is not as I initialy thought. The example used was a single pixel illuminated in the middle of the CCD element as the correct "perfect" image. Then in order to reconstruct the single pixel it cannot be dispersed by the lens any more than the half the minimum resuolution of either axis. This making sense it also follows that the effective resolution of this telescope would then also be one pixel as the neibouring pixels would have lost information. The further from the center the less information that can be used to reconstruct the image.
This effectivily means that the true resolution is always less than the CCD resolution by the amount of deviation/dispersion of any given pixel. Ultimately we can calculate the required error in the reflector that can be compensated for as a function of the original CCD resolution and the amount of usable resolution we want to acheive and the amount of information we are prepared to lose beyond the corrected viewable area.

He even gave me advice on building a suitable mirror with the equipment I have at home and some I need to get..

Thank's for the feedback guy's. Be it positive or negative it's always welcome food for thought. The diversity and depth of knowledge of the members in this forum realy is amazing.



Regards,