Just an idea... It's well-known that the resolution impairment of a ground based astronomical telescope with large opening (>10 inch) is mainly caused by the atmosphere disturbance, which can be modeled as a random process. As a result, the impinging wave front at the telescope objective is no more plane but time-variant and distorted leading to a blurred image instead of an ideal pin-point image of a star. Now it seems to be established, that large ground based telescopes feature adaptive optics in order to equalize the atmosphere disturbance. Here, the crucial component is an adaptive mirror made of elastic materials. On the back side of the mirror hundreds or even thousands of small Piezo cells are detached, which can deform the vicinity of the mirror under the control of an electrical signal in real time. The control signal is generated by some pilot aided equalization algorithms, and the pilot can be a star neighboring the object of interest. This adaptive mirror is supposed to be very expensive to produce. Goggle �adaptive optics� for more details. So, why don't we replace the highly complex mirror by a high speed digital camera and process the image sequence by DSP. Sure, you may introduce some delay, though what matters for stars billions of light years away. lanbaba This message was sent using the Comp.DSP web interface on www.DSPRelated.com
adaptive optics without adptive optics?
Started by ●October 5, 2005
Reply by ●October 5, 20052005-10-05
lanbaba wrote:> Just an idea... > > It's well-known that the resolution impairment of a ground based > astronomical telescope with large opening (>10 inch) is mainly caused by > the atmosphere disturbance, which can be modeled as a random process. As a > result, the impinging wave front at the telescope objective is no more > plane but time-variant and distorted leading to a blurred image instead of > an ideal pin-point image of a star. > > Now it seems to be established, that large ground based telescopes feature > adaptive optics in order to equalize the atmosphere disturbance. Here, the > crucial component is an adaptive mirror made of elastic materials. On the > back side of the mirror hundreds or even thousands of small Piezo cells > are detached, which can deform the vicinity of the mirror under the > control of an electrical signal in real time. The control signal is > generated by some pilot aided equalization algorithms, and the pilot can > be a star neighboring the object of interest. This adaptive mirror is > supposed to be very expensive to produce. Goggle �adaptive optics� for > more details. > > So, why don't we replace the highly complex mirror by a high speed digital > camera and process the image sequence by DSP. Sure, you may introduce some > delay, though what matters for stars billions of light years away.One reason is not that the image dances erratically (we know how to handle that) but different parts of a fixed mirror put the image in different places at the same time. Another reason is that the frequency of photon strikes in the images of faint objects is lower than that of the disturbance. jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 5, 20052005-10-05
lanbaba wrote:> Just an idea... > > It's well-known that the resolution impairment of a ground based > astronomical telescope with large opening (>10 inch) is mainly caused by > the atmosphere disturbance, which can be modeled as a random process. As a > result, the impinging wave front at the telescope objective is no more > plane but time-variant and distorted leading to a blurred image instead of > an ideal pin-point image of a star. > > Now it seems to be established, that large ground based telescopes feature > adaptive optics in order to equalize the atmosphere disturbance. Here, the > crucial component is an adaptive mirror made of elastic materials. On the > back side of the mirror hundreds or even thousands of small Piezo cells > are detached, which can deform the vicinity of the mirror under the > control of an electrical signal in real time. The control signal is > generated by some pilot aided equalization algorithms, and the pilot can > be a star neighboring the object of interest. This adaptive mirror is > supposed to be very expensive to produce. Goggle "adaptive optics" for > more details. > > So, why don't we replace the highly complex mirror by a high speed digital > camera and process the image sequence by DSP. Sure, you may introduce some > delay, though what matters for stars billions of light years away.One way of dealing with this, could be to use an average of a few hundred or thousands of images. The problem is that each image would take a long time to get (exposure times can be very long in astronimical photography), so the efficency of the telescope would plumit, taking only a handful of images where it ought to be able to take hundreds or thousands. The economists might have one or two opinions about such an approach. As for your idea, it might work if stars were the only interesting objects. However, galaxies, star clusters and nebulae, not to mention comets, are at least as interesting objects. These are not point-shaped like the stars, nor is it always possible to find point-shaped features. So an adaptive imaging method might just not have that reference point to use during image processing. One of the key issues of data processing is to be able to separate processing artifacts from features characteristic to the data. It would be very difficult to establish whether some fuzzy feature in the image is an object out there, or merely a glitch in the focusing routine. Such a doubt might kick the feet away under a project or observatory. Rune
Reply by ●October 5, 20052005-10-05
lanbaba wrote:> Just an idea... > > It's well-known that the resolution impairment of a ground based > astronomical telescope with large opening (>10 inch) is mainly caused by > the atmosphere disturbance, which can be modeled as a random process. As a > result, the impinging wave front at the telescope objective is no more > plane but time-variant and distorted leading to a blurred image instead of > an ideal pin-point image of a star. > > Now it seems to be established, that large ground based telescopes feature > adaptive optics in order to equalize the atmosphere disturbance. Here, the > crucial component is an adaptive mirror made of elastic materials. On the > back side of the mirror hundreds or even thousands of small Piezo cells > are detached, which can deform the vicinity of the mirror under the > control of an electrical signal in real time. The control signal is > generated by some pilot aided equalization algorithms, and the pilot can > be a star neighboring the object of interest. This adaptive mirror is > supposed to be very expensive to produce. Goggle �adaptive optics� for > more details. > > So, why don't we replace the highly complex mirror by a high speed digital > camera and process the image sequence by DSP. Sure, you may introduce some > delay, though what matters for stars billions of light years away. > > > lanbaba > > This message was sent using the Comp.DSP web interface on > www.DSPRelated.comMirrors form images coherently, that is, with both the phase (actually flight delay) and ampltitude. A camera gives you an image, the phase information is lost. There are techniques that work on sequences of images but the processing is incoherent.
Reply by ●October 5, 20052005-10-05
Stan Pawlukiewicz wrote:> lanbaba wrote: > >> Just an idea... >> >> It's well-known that the resolution impairment of a ground based >> astronomical telescope with large opening (>10 inch) is mainly caused by >> the atmosphere disturbance, which can be modeled as a random process. >> As a >> result, the impinging wave front at the telescope objective is no more >> plane but time-variant and distorted leading to a blurred image >> instead of >> an ideal pin-point image of a star. >> >> Now it seems to be established, that large ground based telescopes >> feature >> adaptive optics in order to equalize the atmosphere disturbance. Here, >> the >> crucial component is an adaptive mirror made of elastic materials. On the >> back side of the mirror hundreds or even thousands of small Piezo cells >> are detached, which can deform the vicinity of the mirror under the >> control of an electrical signal in real time. The control signal is >> generated by some pilot aided equalization algorithms, and the pilot can >> be a star neighboring the object of interest. This adaptive mirror is >> supposed to be very expensive to produce. Goggle �adaptive optics� for >> more details. >> So, why don't we replace the highly complex mirror by a high speed >> digital >> camera and process the image sequence by DSP. Sure, you may introduce >> some >> delay, though what matters for stars billions of light years away. >> >> lanbaba >> >> This message was sent using the Comp.DSP web interface on >> www.DSPRelated.com > > > Mirrors form images coherently, that is, with both the phase (actually > flight delay) and ampltitude. A camera gives you an image, the phase > information is lost. There are techniques that work on sequences of > images but the processing is incoherent.Astronomers are also interested in the light spectrum. Using adaptive optices they can coherently collect as many photons as possible and feed them all to a difraction grating. Using your proposed technique this would be much more complicated.
Reply by ●October 6, 20052005-10-06
"lanbaba" <lanbaba@gmx.ch> wrote in message news:YPadnSlOe_G6Vd7eRVn-sQ@giganews.com...> So, why don't we replace the highly complex mirror by a high speed digital > camera and process the image sequence by DSP. Sure, you may introduce some > delay, though what matters for stars billions of light years away.Mostly because high-speed digital cameras need a lot of light, and astronomers mostly take pictures of dim things. On a professional telescope, a single exposure of a very faint object can take hours. -- Matt
Reply by ●October 6, 20052005-10-06
Matt Timmermans wrote:> "lanbaba" <lanbaba@gmx.ch> wrote in message > news:YPadnSlOe_G6Vd7eRVn-sQ@giganews.com... > >>So, why don't we replace the highly complex mirror by a high speed digital >>camera and process the image sequence by DSP. Sure, you may introduce some >>delay, though what matters for stars billions of light years away. > > > Mostly because high-speed digital cameras need a lot of light, and > astronomers mostly take pictures of dim things. On a professional > telescope, a single exposure of a very faint object can take hours. > >Lanbaba, to add to what Matt says, during this time the atmosphere will have smeared the image somewhat, even at the best observing site. Signal processing can improve the definition and contrast of a bad image, but it does an even better job of improving a good image. Where possible, it is best to use adaptive optics initially, to generate the best possible image as the first step, and then to apply a measured amount of processing to improve the image further. In some telescopes the adaptive optics mechanism serves to keep the mirror (or mirrors) precisely in shape (and aligned) to compensate for flexing and expansion in the mirror and its supports. Unfortunately, in the majority of cases there is no star within the field of interest that is bright enough to be used as a reference for second-by second correction of the optics, and so it is not possible to compensate for atmospheric turbulence during a long exposure. Regards, John
Reply by ●October 6, 20052005-10-06
Matt Timmermans wrote:> "lanbaba" <lanbaba@gmx.ch> wrote in message > news:YPadnSlOe_G6Vd7eRVn-sQ@giganews.com... > >>So, why don't we replace the highly complex mirror by a high speed digital >>camera and process the image sequence by DSP. Sure, you may introduce some >>delay, though what matters for stars billions of light years away. > > > Mostly because high-speed digital cameras need a lot of light, and > astronomers mostly take pictures of dim things. On a professional > telescope, a single exposure of a very faint object can take hours.Actually, astronomers use image intensifiers and CCD arrays, both at cryogenic temperatures to reduce noise. Despite the cold, exposures of only a few minutes are common, the accumulated noise becoming dominant with longer ones. Averaging many such pictures seems to be the better way, for reasons I can guess but can't verify. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●October 7, 20052005-10-07
"Matt Timmermans" <mt0000@sympatico.nospam-remove.ca> wrote in message news:vGj1f.3384$R4.526640@news20.bellglobal.com...> > "lanbaba" <lanbaba@gmx.ch> wrote in message > news:YPadnSlOe_G6Vd7eRVn-sQ@giganews.com... >> So, why don't we replace the highly complex mirror by a high speed digital >> camera and process the image sequence by DSP. Sure, you may introduce some >> delay, though what matters for stars billions of light years away. > > Mostly because high-speed digital cameras need a lot of light, and astronomers > mostly take pictures of dim things. On a professional telescope, a single > exposure of a very faint object can take hours.Just guessing here, but I would think that on the big telescopes fancy enough to use adaptive optics, cost isn't really a big issue. And compared with the cost of everything else, I don't think the adaptive part is relatively all that much (I read about one that's going to take about 10 years to complete--that can't be cheap!).
Reply by ●October 7, 20052005-10-07
>lanbaba wrote: >> Just an idea... >> >> It's well-known that the resolution impairment of a ground based >> astronomical telescope with large opening (>10 inch) is mainly causedby>> the atmosphere disturbance, which can be modeled as a random process.As a>> result, the impinging wave front at the telescope objective is no more >> plane but time-variant and distorted leading to a blurred image insteadof>> an ideal pin-point image of a star. >> >> Now it seems to be established, that large ground based telescopesfeature>> adaptive optics in order to equalize the atmosphere disturbance. Here,the>> crucial component is an adaptive mirror made of elastic materials. Onthe>> back side of the mirror hundreds or even thousands of small Piezocells>> are detached, which can deform the vicinity of the mirror under the >> control of an electrical signal in real time. The control signal is >> generated by some pilot aided equalization algorithms, and the pilotcan>> be a star neighboring the object of interest. This adaptive mirror is >> supposed to be very expensive to produce. Goggle �adaptive optics� for >> more details. >> >> So, why don't we replace the highly complex mirror by a high speeddigital>> camera and process the image sequence by DSP. Sure, you may introducesome>> delay, though what matters for stars billions of light years away. >> >> >> lanbaba >> >> This message was sent using the Comp.DSP web interface on >> www.DSPRelated.com > >Mirrors form images coherently, that is, with both the phase (actually >flight delay) and ampltitude. A camera gives you an image, the phase >information is lost. There are techniques that work on sequences of >images but the processing is incoherent. >An answer to one or another: I think the text above can lead to a reasonalble argument. The image processing what I mean here to replace the adaptive mirror is not to enhance the image quality by adjusting contrast or so, nor it is a simple diversity combining by averaging many noise corrupted images. The only goal here is to compensate for the atmosphere turbulance by analyse CCD images. In terms of communications it is some kind of channel equalization. If no suitable star in the field of view is available as pilot one can use a laser beam as active pilot. So, pilot is not a problem. The difference of hardware adaptive optics with adaptive mirrors and my software solution is rather big when I thinked over about it, and this should be exactly the reason why the latter cannot do the equalization job. In the hardware solution, the adaptive optics is something like a hidden Markov process. You have observation which is the impinging light wave front at the objective. Note the image resulting at the CCD camera is NOT observation by only an estimation of the object of interest. By comparing this estimation with the reference signal one can deduce a control signal to change the state space which is the state of all Piezo cells on the adaptive mirror. Consequently, the input signal or the observation can be equalized. Most communication engineers should be familiar with this... In the software solution, the observation is lost. What lost is lost. Therefore, no equalization is possible. Conclusion: the hardware implementation with adaptive mirrors is the only feasable method to compensate for the atposphere turbulance untill a method to save the impinging wave front is invented. Lanbaba This message was sent using the Comp.DSP web interface on www.DSPRelated.com
