Hola,
I beg immediately the pardon of the mathematicians, the following message is only to give an approximative idea of how it works.
The wavefront of the light emitted by a star is distorted by atmospheric cells which shift slightly the light emitted and produces their own image of the incident light. Instead of having only one perfect image of the star we get a heap of points : the speckles (the phenomenon is far more complicated -ie. it introduces interferences- nevertheless the result is identical when we look through the eyepiece
. Each speckle is slightly different from the others, they move very fast around the mean position of the star. Their main quality is that they approximate closely the diffraction image of the objective -> the resolving power is kept.
The number of speckles depends on the ratio between the diameter of the telescope and the size of the atmospheric cells causing the distorsion. Increasing the diameter of the telescope or decreasing the size of the cells increases the number of speckles. If the mean size of the cells is high in regard of the telescope aperture, we get an perfect image of the star. It's the reason why the stars seems far more beautiful in small telescope. Its low resolving power masks the small number of speckle in the Airy image. In larger diameter the stars explodes into speckles.
The sequence of schematic images shows from left to right :
1) A single star, perfect seeing
2) A pattern of speckles of a single star
3) A close double star
4) The melted pattern of speckles of the double star on a large telescope. Both stars have the same pattern. It occurs when their respective lights cross the same cells of atmospheric distortion. It works only for quite close stars !
5) What is seen on the image at the telescope. It's difficult to see that it's a double, isn't it ?
6) The autocorrelogram on the same after Reduc processing
explic.jpg
As explained above every cells produced its own image of the double star, the resolving power is kept. If we measure all the distances between the couples of speckles and draw the histogramm of their distribution, we will see that one distance is repeated far more than others. It's the distance between the two components. If you don't believe me, you can count
The information is there. It remains only to extract it with an appropriate tool. Autocorrelation, intercorrelation and DVA works in similar ways and uses FFT for the analysis (the complete mathematical process is largely described on the web). In all case the result is an image showing a series of peak. The three central peaks are the representation of our double star. This symetrical shape is caused by the mathematical process and the measurement is done between the two external peaks.
With autocorrelation an ambiguity of 180° appears and must resolved by other means (not really a problem). Intercorrelation and DVA gives a direction when the stars are of enough different magnitudes.
Alas all isn't so easy
The speckles are very dim and must be as numerous as possible -> a (very) large telescope is needed
The speckle pattern move very fast -> a very fast camera is needed
The speckles are dim -> a very sensitive camera is needed
The signal/noise ratio is very low -> many images are needed
The speckles are blurred by chromatic effects -> it's better to use filters ... but the speckles become dimmer
... and so on.
Finally, very few bright doubles can be resolved by speckle interferometry with classical amateur instruments. The interest of measuring these stars is low because they are easily and regularly scanned by the professionals. In order to work on dim stars and to observe as often as possible, I use a method just between spatial imagery and speckle interferometry. Exposure times are a bit too long to freeze the speckles but short enough to avoid a complete blurring of the image (what I call super-speckles) and the reduction processing mimics the speckle interferometry. You can see the result on the animations above in the thread or below with a sequence taken on Gamma Virgo last spring.
Only the last image shows clearly the two components nevertheless we see that useful information is carried by the other images. The autocorrelation permits the measurement.
gamvirgo.jpg
gam_vir_tout.jpg
As cited by MigL, Argyle's book contains a very good introduction and many other useful things.
Many links on the web for every levels can be found by typing : speckle interferometry in Google.
Very interesting articles can be found at the ADS :
http://adsabs.harvard.edu/abstract_service.htmlType speckle interferometry in the second textbox then Send Query. Articles from Hartkopf, Mason, Mc Alister published around 1990 explain many things.
Thank you Nachote for the link to the thesis. I was wondering if something could be done with the Audine in drift-scan mode, it seems that it's the case . It's worth a try
Un abrazo,
Florent
No tiene los permisos requeridos para ver los archivos adjuntos a este mensaje.