Mark, I think you are just talking about unit transform, or dimensional analysis? Am I correct? If this is the case, I think the astrophysicist have that one covered :-)
Though of course I could be misunderstanding you. Joel On Wed, Feb 27, 2013 at 4:25 PM, MARK BAKER <[email protected]> wrote: > > > That is really Fascinating, may be there might be a underlying > representation of > this tho not ordered (I think that is what you meant by Theoretical models) > > I will do some googling to see if I can perhaps find some commonly used > units in astronomy, as well I will do some book search, It seems like > something like this, should be beneficial for both astronomy and molecular > physics, I will see if I can find a list of units to work with... > > maybe this can help to find the dimension of Dark Matter... > my guess is it would be a dimension integrated of [mass]^-6 .. [mass]^-11 > > Cheers > -Mark > > ------------------------------ > *From:* Jarle Brinchmann <[email protected]> > *To:* Craig DeForest <[email protected]> > *Cc:* MARK BAKER <[email protected]>; """[email protected]""" < > [email protected]> > *Sent:* Wednesday, February 27, 2013 1:47 PM > > *Subject:* Re: [Perldl] radial power spectrum > > To follow up on what Craig says, what you are describing, Mark, is indeed > something we do regularly in extra-galactic astronomy but the > transformation of image values has to go through a mapping which is > ultimately based on theoretical models. > > Thus if you have flux with an effective wavelength of l1 and l2, say, > there are functions that you can apply that can map > > G[ f(l1), f(l2)] -> Mass of stars > > for instance. There are a number of technical problems with this, however, > not least that the function is often multi-valued - for a given x, y there > are a number of possible G(x, y). These problems are usually even more > serious when you consider other possible physical quantities you might want > to extract from the images such as mean ages or metal content. However it > is a major area of research in astronomy and is indeed very valuable - if > you want an overview you can do worse than consult the very recent review > article on the subject: http://esoads.eso.org/abs/2013arXiv1301.7095C. > > Other wavelengths than the optical provide other information but the > general idea is much the same although there are quantities that are more > reliably estimated from the data. And expanding the spectral dimension so > that you do not have several colours, but rather have a spectrum in each > spatial pixel provides even more of an improvement and optimally exploiting > this dimension is still work in progress I would say, at least for distant > galaxies. > > Cheers, > Jarle. > > > > On 27 Feb 2013, at 22:34, Craig DeForest wrote: > > > Well, there are a lot of things different people have gotten from images > - but, fundamentally, all that astronomers get from images is brightness > versus arrival angle and wavelength. Spectral power density, in the context > of an image, is a map of how much brightness exists at different spatial > scales in the image -- i.e. if there are a lot of big, diffuse things, > there will be a lot of spectral power at low spatial frequencies (big > spatial scales), and not much at high spatial frequencies (small spatial > scales). > > > > You see astronomers talking about mass of this and mass of that, but > those masses are inferred using methods that are far more sophisticated and > specialized than PDL::Transform can accomplish. > > > > > > > > On Feb 27, 2013, at 2:26 PM, MARK BAKER <[email protected]> wrote: > > > >> Hey Craig, > >> > >> [I'm not sure what you mean by, for example, "now we make a > transformation for Mass to any power".] > >> > >> so based on the position of each individual pixel and it's color > compared to other pixels and there color... > >> we should be able to map transformations of different Dimensions , yet > to do this simply we have to have > >> known mapping transformations of known Dimensions, then we should be > able to derive > >> mapping transformations of any dimension. > >> > >> This Idea would change the pixel colors and pixel position to highlight > different dimensional values. > >> > >> if you can send me a list of commonly used Astronomical dimensions that > have been measured accurately > >> from a image , I should be able to show you a better explanation > Mathematically > >> > >> just send me a list of units like the "Spectral Power Density" used in > Astronomical transformations > >> > >> > >> Cheers > >> > >> -Mark > >> > >> > >> From: Craig DeForest <[email protected]> > >> To: MARK BAKER <[email protected]> > >> Cc: Craig DeForest <[email protected]>; John Lapeyre < > [email protected]>; ""[email protected]"" < > [email protected]> > >> Sent: Wednesday, February 27, 2013 1:04 PM > >> Subject: Re: [Perldl] radial power spectrum > >> > >> Mark, I'm not sure what you're getting at here. The Transform module > only does coordinate transformations on data sets. It modifies vectors or > images so that the components of the vector, or pixel indices of the image, > have a different geometry than they originally did. > >> > >> An image, for example, is a collection of values taken at a regular > grid of positions (X,Y): one pixel index is proportional to X position in > the image, and the other pixel index is proportional to Y position in the > image. With Transform::map, you can resample the image so that the pixel > indices are proportional to some other parameter (like distance from a > particular point, or angle *around* that point). > >> > >> I'm not sure what you mean by, for example, "now we make a > transformation for Mass to any power". > >> > >> Cheers, > >> Craig > >> > >> > >> On Feb 27, 2013, at 1:44 PM, MARK BAKER <[email protected]> wrote: > >> > >>> > >>> you Might have hinted on to something very big here ... > >>> I thought about this for a while , and here is what I have > >>> if you have a know dimension that can be found threw image processing > >>> say ([mass][length]^3[time]^-4[current]^-2) and if you can find some > other dimensions > >>> then you might be able to derive a image transform for each dimension > >>> > >>> as a example voltage/resistance = current so now you have the I > (current) dimension > >>> resistance * capacitance = time so now we have our T (time) > dimension > >>> speed of light / frequency = wave length so now we have our L > (length)) dimension (1/time = frequency) > >>> and now voltage * L^-2*T^3*I^1 = M so now we have our Mass dimension > >>> > >>> now we make a transformation for Mass to any power > >>> now we make a transformation for Length to any power > >>> now we make a transformation for Time to any power > >>> now we make a transformation for current to any power > >>> > >>> by mixing those dimensions now now we can process a value for any > >>> unit Dimension like the Power spectral density = [Mass]*[Length]^2 * > [Time]-2 > >>> > >>> with this Idea you can calculate all 194481 value in string theory of > the image ... > >>> > >>> if you can find a few different transformations and can send them to me > >>> I would be happy to try to help build a multi-dimensional imaging > engine ... > >>> > >>> Perfect Blessing's > >>> -Mark > >>> > >>> "sometimes I think perl is alive". > >>> > >>> > >>> From: John Lapeyre <[email protected]> > >>> To: Craig DeForest <[email protected]>; " > [email protected]" <[email protected]> > >>> Sent: Saturday, February 23, 2013 1:26 PM > >>> Subject: Re: [Perldl] radial power spectrum > >>> > >>> > >>> Awesome. Thanks. Have fun! > >>> > >>> On 02/23/2013 09:52 PM, Craig DeForest wrote: > >>> > I fft rhem use PDL::Transform > >>> for the radial part. Periodic boundaries are your friend. Sorry > >>> for brief - on ski lift. > >>> > >>> > >>> > > >>> > >>> > >>> > (Mobile) > >>> > >>> > >>> > > >>> > >>> > >>> > > >>> > >>> > >>> > On Feb 23, 2013, at 11:40 AM, John Lapeyre > >>> > >>> <[email protected]> wrote: > >>> > >>> > > >>> > >>> > >>> >> Greetings, > >>> > >>> > >>> >> > >>> > >>> > >>> >> I want to compute the power spectral density of an image, > >>> and then > >>> > >>> > >>> >> integrate over the azimuth to get a radial (in wavelengh) > >>> spectral > >>> > >>> > >>> >> density. I wonder if anyone has code to do this ? I am > >>> trying to cook > >>> > >>> > >>> >> up something with rvals, and whichND, but I don't want to > >>> waste time > >>> > >>> > >>> >> if it is already coded. > >>> > >>> > >>> >> > >>> > >>> > >>> >> Thanks, > >>> > >>> > >>> >> John > >>> > >>> > >>> >> > >>> > >>> > >>> >> > >>> > >>> > >>> >> _______________________________________________ > >>> > >>> > >>> >> Perldl mailing list > >>> > >>> > >>> >> > >>> [email protected] > >>> > >>> >> > >>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > >>> > >>> >> > >>> > >>> > >>> > >>> > >>> _______________________________________________ > >>> Perldl mailing list > >>> [email protected] > >>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > >>> > >>> > >>> _______________________________________________ > >>> Perldl mailing list > >>> [email protected] > >>> http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > >> > >> > >> > > > > _______________________________________________ > > Perldl mailing list > > [email protected] > > http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > > > > > _______________________________________________ > Perldl mailing list > [email protected] > http://mailman.jach.hawaii.edu/mailman/listinfo/perldl > >
_______________________________________________ Perldl mailing list [email protected] http://mailman.jach.hawaii.edu/mailman/listinfo/perldl
