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
