Creation of high-resolution lunar / planetary images ( version 3.3)
By Peter Wellmann

• Part 0: mount and optics
• Part 1: maximum achievable resolution
• Part 2: choice of camera
• Part 3: adjustment of the camera
• Part 4: preparation and implementation of a consumption
• Part 5: Normal image processing
• Part 6: sharpening images
• Part 7: image enhancement
• Part 8: Summary of Results
• Part 9: filter and infrared photography

Preface : Version 2 of our tutorials has been greatly expanded in scope, with more precise determination of the focal length of our telescope , it came with some previously published figures to minor adjustments. Over the years, the technique for creating images in our Moon has refined , and the devices were better. In consequence emerged Moon images of very good quality, for the type of telescope used by us ( Schmidt Cassegrain ) and the available aperture ( 30cm) almost are at the limit of what is theoretically possible. Our images landed in books, on Astro- CD , as "Picture of the Day" in the U.S., and magazines have been printed in known astronomy. Since it is not easy to produce such images , we have put together what seems important to us in this context. Almost all of the proposed methods are also suitable for certain areas of the planetary photography. So if images as the following examples 1 Terminator 2 Rupes Recta 3 Apennines 4 Saturn 5 Mars 6 Jupiter IrRGB , is safe interested in this information . All statements in this " tutorial" come from the practice of our recordings , formulas are not derived from unmanageable theoretical approaches but from pictures moon recognized photographer , and merely describe their practice. Who disagrees that must be like !


Part 0 : Mount and Optics :

The Mount: On the mount there are no special requirements. For the recording of still images even a polar alignment is unnecessary, the mount may , however, produce in tracking (and possibly also in the correction ) no vibrations. This can be checked most easily by the observation of a star at extremely high magnification and a quiet air. To counter- check to turn off the tracking , the star then moved completely calm and straight from the field , so a possibly previously existing tremor was caused by the mount , and not by poor seeing. A hand controller for fine movement and for operating a Elektrofokussierers would be very helpful. If mosaics are created from multiple images , the mount must be relatively well aligned polar, so do not twist at the longer recording time for the series the images against each other.

Optics : High resolution photography is ( perfect optics and camera required) a pure opening question. Tell me which aperture your telescope, and I tell you what resolution you reach. Crucial to the maximum resolution is also in planar images , the diameter of the central diffraction disk ( Airy disk) that provides a look from a light spot , because even a planar image of individual, overlapping pixels builds. Default mirror obstructed (obstruction = partial coverage of the primary mirror by the secondary mirror located in the beam path ) are at high resolution compared to refractors rather at an advantage because they can produce significantly smaller Airy disks, and thus provide higher resolution. Although they have more light in the diffraction rings , and therefore a lower contrast than a refractor (ie a worse visual sharpness ), but that is very easy to compensate photographically in image processing. Way obstructed systems the nose of this, also in the separation of double stars in front.

The poorer reputation of the mirror is based essentially on the fact that apochromatic (ie pure color ) refractors large opening are extremely expensive , and therefore in principle be of high quality (who can be a cucumber attach already for several thousand or even several thousand euro ) , mirror however, almost all come from more or less cheaper mass production , and also come in a correspondingly reduced average quality on the market. While there are individual devices (rightly ) good reputation , such as Celestron C9 , 25 with its legendary quality , but it remains an amateur not spared any mirror optics with a 98 % Strehl Protocol ( Strehl 100% = perfect optics ), paper is patient, even to the suitability for high-resolution photography testing.

Since the test ( destroyed Seeing = air turbulence , bad seeing the image ) in the sky because of the usual average seeing conditions is problematic , leaving only the terrestrial version. One considers the minutely and meticulously collimated telescope ( this adjustment is extremely important who since has no sufficient experience should necessarily an expert data sheets ) in very calm air ( if necessary also a day) at least 100m distant , extremely fine- membered object ( eg small scratches on the plate of a roof covering hairline cracks in the plaster of a fireplace , but better at night fine structures like the wings or legs of insects that depend on the glasses of many streetlights ) with a good eyepiece. Are these structures emotionally good sharp up to a magnification

• Simple Lens Diameter in mm , so has the appearance of significant deficiencies and is not beugunsbegrenzt ( Strehl below 80 % ) for high-resolution photography unusable
  
  
• Double Lens Diameter in mm, so the look is not entirely free of defects but beugunsbegrenzt ( Strehl above 80% ) for high-resolution photography useful
  
  
• Triple Lens Diameter in mm, so has the look except minor roughness etc. no defects ( Strehl above 95 %), ideal to use for high-resolution photography.
  

It may surprise some that we apply the last category , which is often a complaint for refractors on the mirror, but as shown above provide perfect mirror photographically theoretically at least as sharp images as refractors , and our tests on dozens of telescopes have this procedure confirmed. Thus provides the 30cm SC optics used by us at the moment to far "empty" in the area of ​​magnification in at 850fach on my terrestrial standard obje ct images absolutely blameless , and that, although the optics in very detailed interferometric test quite considerable roughness and other small error shows. Finally, it should again be noted that such increases can not be checked in the sky because the air will in our regions ( almost) never quiet enough ! Our experience shows ( unfortunately) that only a small portion of the products of mass production reaches For the third stage of the above scale. Meets the optics only the first stage , we should immediately return it to the dealer!


Part 1: Maximum resolution at the moon / planet images

point figure: Unfortunately, a look a bright spot (eg a star ) does not represent point-like. The diffraction of light to the round aperture of the telescope with a diameter D leads to a bright slices ( Airy disc) surrounded by a plurality of ideally only faintly visible light rings. Critical to the maximum resolution in the image is the diameter of the optics of the Airy disks, the smaller it is, the better the resolution of the optics. The diameter d of the Airy disks applies a function of the wavelength λ of the light , and the opening ratio f ( " aperture" ) of the formula d = 2.5 ? F. For a lens with 30cm diameter at 3m focal length (f10 ) gives eg for 550nm wavelength , a diameter of about 14 micrometers ( For professionals : That is about twice the half width of the PSF).

resolution at point figure: " resolution " is in the context of optical telescopes of diameter D is a well-defined technical term and is regulated in the form of Rayligh or Dawes criterion. It is about the question " At what angle distance a the picture of two identical light dots is just the image of a single bright point distinct" ? The Dawes criterion ( for small angles a = λ / D) assumes that two Airy disk in the center distance of less than half its diameter just to be recognized as a superposition of two slices still even. The Rayligh criterion ( for small angles a = 1.22 λ / D) is the center of a Scheibchens in the first dark ring around the second slice , is by a factor of 1.22 less optimistic , and calls for a by this factor greater distance. After Dawes optics with 10cm opening provides rounded at about 550nm wavelength of light 1.2 arc sec resolution. The resolution increases in proportion to the diameter of the optical system, a 30cm - optics solves therefore three times better on , and reached 0.4 " (exactly 0.38 "). This means that at an angular distance of 0.38 " two equally bright points just as two points can be detected, although their images overlap for more than half.

resolution at surfaces : Based on above definition now the question: " At what diameter a structure with high contrast on the moon is just visible " ? To this end, there is no clear definition , there but also a planar image of single pixels composed for the Dawes criterion applies because of the undisturbed superposition of waves , it can be assumed that this criterion also approximately the resolution in planar image ( moon , planets) describes. But I want to make at this point no complicated theoretical approaches to the derivation of a formula for the resolution in planar images, which then in turn lead to endless discussions. Therefore, I will limit myself from the world's best amateur photos and sharp passages from their own images by comparison with Moon NASA images to determine the finest visible features and to take the limits thus obtained in a formula. Such a pure description of the practice requires no principled justification , errors due to the uncertainty in determining the crater diameters are in the single digits.

An example of the procedure: Among other things, we have small sharp areas in our Clavius ​​image wanted. These areas we have compared with a high -resolution Clementine Recording ( Map-A-Planet/USGS ). As a benchmark , we have applied the inner diameter of the crater Clavius ​​D with 27km. Because of not well-defined boundaries of the outer diameter of the crater walls all were based on the same edges of the walls.

To understand the following numbers is important to note that crater diameter must of course be determined from the NASA images and not (!) Of the telescopic images because below a certain limit , the diameter of the telescope picture not by the real crater diameter is determined , but assumes a nearly constant value, which is determined by the PSF (Point spread Function ) of the optical system (for more details see below). From the examples resulted in compliance with the lunar distance at the time of the shooting the following result ( mileage applies for a mean lunar distance ) :

• crater diameter > 1.6 km and 0.84 " : craters are mapped structured , the better , the more the value above the 0.84 " is. ( The value is almost twice worse than the Rayligh criterion). Generalized for any very good telescopes in the opening D in meters applies to the crater diameter a in arc seconds: a = 0.25 / D.
  
  
• crater diameter > 0.75 km or 0.42 " : craters are just clean and contrasty detected , the better , the more the value above the 0.42 " is. The picture is already significantly influenced by the PSF. ( The value is between Dawes and Rayligh criterion). Generalized for arbitrary telescopes applies to a very good look: a = 0.13 / D.
  
  
• crater diameter > 0.56 km and 0.32 " : The craters are only more or less indicated and all appear under the same , intended by the PSF diameter , the visibility is even better, the further the value is above the 0 , 32 " is. 0.32 " is the contrast Zero , that is, the crater is no longer detectable even with the best intentions . The value is better than the Dawes criterion generalized for arbitrary telescopes applies to a very good look: a = 0.1 / D.

Conclusion: Under the best conditions with absolutely perfect optics , the best moon photographers have occasionally marginally achieved world better values. As an example, a Plato - recording with a C14 ( 35cm opening) could be considered by Peach , which was trimmed for maximum resolution , and about 50 small craters in Plato level down shows to 500m in diameter. Thus the question of the resolution for good amateur telescopes is answered : A relatively reliable mean value of the angular resolution for the safe detectability of fine structures is represented by the formula of the second discussed above , if the formula a = 0.13 / D given , this value is the mean of Rayligh and Dawes criterion. Under the best conditions are still substantially finer structures " detected " but no longer geometrically " mapped ". This is especially true for long drawn structures (Moon grooves , Encke division in Saturn) , which may be visible below the above indicated limits on pictures yet also. Owner very perfect optics can expect , at best, with again slightly better values, objects with low contrast range is , however, reduce the values.

In general, large optics will have more trouble to achieve the above limits, because it is much harder to make a big appearance with the required quality. The contrast provides the optics is also important. If it is high ( more expensive APO refractor ) , a borderline resolution will naturally be reached sooner than with an inexpensive window glass with SC -Schmidt plate. Each user must so arrange his appearance according to these criteria themselves and then decide which of the above-mentioned limits he achieve so......

• rule of thumb for a very good look at high contrast object : Under the best conditions provide Dawes and Rayligh criterion not only the point image but also in the planar images for the detectability of small details a realistic value. Corresponding rule of thumb for 550nm and a in arcseconds and D in meters a = 0.13 / D. For the geometric resolution of details about you need twice the Dawes criterion borderline " detecting " of structures can be achieved even well below the limit of Dawes to about a = 0.1 / D. For the detection of grooves and similar structures even enough already about 0.6 times that of the Dawes criterion.
  
  

A 20cm optics will show craters, which are only 0.2 arcsec larger than a 30cm optics. But be careful ! Between these values ​​lie at the moon all the worlds! Doubling the resolution does not lead to a doubling , but to a much larger increase in the visible structures because on the moon in many regions the number of objects with decreasing size is increasing rapidly ! One can easily verify on an Apollo image of Ptolemy or another crater- rich region. The moon gets up a brand new image and it's worth it to switch from 20cm to 30cm diameter optics. 35-40 cm optics are then again , a whole class better, and the 30cm optics is no chance for high-resolution images is then sufficient but seeing only a few days a year....

What do you really see? Now a more important note ! With the above considerations, it comes when a structure on the moon or a planet just in the visual or photographic image is " detectable ". At the limit , however, this in no way gives a geometric image of these small structures , so they are not really " resolved" , but only " detected ". In absolute borderline case is , for example, in our 30cm optics, the angular size just detectable structures about 0.32 " to 0.42 ", which participated in the Figure Airy disk but have a diameter of up to 0.9 ". Which is two to three times as large as the actual image structure whose shape is thus determined largely by the shape of the Airy disk ( or the PSF). When approaching the limit of detectability, ie the " geometric " figure comes increasingly into an "artifact" with a diameter about whose appearance is largely determined by the PSF. In practice, the smallest weakly indicated craters have a high resolution moon picture all nearly the same size , and have no clear conclusion as to the actual size of the crater on the moon . This also explains the often false information about the " Resolution" of images: one measures on the PC screen expansion one small , extremely low-contrast crater of 5 pixels and determines the diameter , so could the actual diameter on the moon be considerably less ! This is also true for other structures that Encke division in Saturn's ring is much wider imaged by small optics for the same reason ( better: " detected " ) than expected.



What this looks like in practice , the comparison of two images of the crater Plato , which were made ​​with telescopes of very different opening (aperture). The lower Image clearly shows the large opening different sized craters in Plato level , the double crater above are completely isolated , contours are sharp. The upper , excellent for small used opening photo shows exactly the features that can be expected at full Ausreizen the theoretical limit in an image : the small crater in Plato appear through the PSF bloated and almost the same size, the double crater above are therefore not with space separated. There are also rounded , co-determined by the PSF edges at all other small structures ( not to be confused with the noise).

The image of the small opening by the way shows a strong noise ( grainy appearance ) which has a much smaller structure than the smallest real image structures. With a noise filter so you could remove this noise quite effectively out of the picture , a consequence of the large aperture ratio ( f46 ) with which the picture was taken. In the picture below (taken with f21 ) would sound the same size as in the picture above , but could not be removed because the smallest real details of the same magnitude as the noise would be , a fact that is still talked about later.



Part 2: Selection of the CCD camera at the moon / planet images

Monochrome or color? A color camera conventional design has a disadvantage compared to a monochrome camera. About the pixels of the CCD is a mask of small color filters in repeating groups of 2x2 = 4 filters. Of these two filters are diagonally opposite green, and one each red or blue ( Bayer mask). From the intensity values ​​of the thus-exposed pixels ( raw image ) of the computer ( or the built- in camera electronics ) must calculate the individual pixels of the color image according to complicated algorithms. Both the brightness and the color is obtained by interpolation. This is of course not without problems , and succeed depending on the used algorithm (not all cameras are there the same) times better , sometimes bad. In any case, of course, results in a loss of sharpness compared to a corresponding monochrome camera. The resulting color interpolation , where appropriate, color fringes , which are not identical with the adjacent colors. The picture below shows the problem from the photograph of a razor - edge in yellow lamplight. That there is no sharp line and in the monochrome image is determined by the diffraction of light at the opening of the optics. In the present case , the actual edge of Razor blade in the middle between the red lines , the theory requires a broadening to about three pixel widths , what good is confirmed in the image. Important here is the broadening of the edge in the color camera , resulting in a larger "effective pixel size " results than with the monochrome camera. This here factor of about 1.3 depends on the position of the edge to the pixel grid of the camera , and reduces the sharpness accordingly. (When using very narrow-band color filters , this factor can increase to a maximum of 2 when only the blue or only the red pixels are addressed). Note we now also the significant loss of light intensity through the filters of the Bayer mask , so to speak following points against the use of a color camera :

• Reduced sensitivity to light
  
  
• Reduced sharpness
  
  
• color interpolation at the edges
  
  

So we will not use color camera for moon photography. The situation is different in the planetary photography because there on color can not be waiv ed. Despite the
above -mentioned disadvantages can be produced , as we will see in examples below
with color cameras beautiful pictures. However, who is dependent on the best image
sharpness and optimum color representation (which are specially users of small optics
) that must access ( for color values) to a combination of monochrome camera like
the DMK21 ( for brightness values ​​) and color camera like the toUcam , or with
make a monochrome camera ( eg red , green, blue) color filter , three images , and
process it to an RGB color image in the computer. Some information to follow below!









CCD or CMOS? On the market are increasingly cameras that use the ultra low-cost
CMOS sensors instead of the expensive traditional CCD sensors. Besides the price
advantage also other points (such as almost no electronics required , low power
consumption, etc. ) speak for CMOS. In professional cameras , this sensor type is
also very excellent characteristics , while these sensors are developed at high
cost special parts that are not available to the general market , of course, for
reasons of competition available. In the inexpensive cameras that are eligible for
astronomy often installed sensors that are not nearly the quality of a "right" Reach
the CCD chip. This is now not mean that all offered astronomical CMOS cameras are
problematic , it should only be a hint that accurate information on the quality
of the camera is at least at the time in 2009 before buying essential! The picture
below shows the comparison between a typical inexpensive CCD sensor and a corresponding
CMOS sensor under identical recording conditions.




width="503" height="102">




A year later , in late 2010, already saw the matter differently. At this time there
were already CMOS sensors significantly better quality. There was particular interest
are the Tatsche that these sensors had some very small pixels. At 2.2 microns pixel
size , you can already do without the focal length with a Barlow with many telescopes
, and work directly in the focus even with high-resolution photos. This is a huge
advantage. We had purchased a camera with such a sensor , and therefore the CMOS
image bottom right bottom left and made ​​to compare the CCD image by superimposition
of 100 frames of a video. The pictures show a small part of a distant street lamp
taken with an ETX telescope through a window glazing throughout , and are not primarily
intended as a sharpness test , but to compare the sensitivity to light. Both cameras
have been set to the same noise in the single images , the focal length of the telescope
is in the CMOS camera is halved by means of a reducer to achieve approximately the
same magnification. To perform an exact comparison of the image scale was corrected
by Nachvergrößern the DMK image even slightly by a factor of 1.17. Had this correction
performed optically , as by the exposure time of the DMK image by a factor of 1:17
² = 1.37 would become longer, which would have led to a time of 1/8s (instead of
1/11s ). Comparing the times of 1/24s with the so- corrected time 1/8s , so you
recognize a factor of 3 in favor of the CMOS camera. The pixels of the CMOS camera
are both in absolute terms somewhat less sensitive than the pixels of the CCD camera
, but by smaller and therefore much lighter image onto a chip CMOS camera can be
reached at the telescope ultimately a gain in sensitivity compared to the previous
lunar reference camera DMK41 , and is even somewhat better than the previous planetary
reference camera DMK21.









So immediately change camera? Unfortunately, the tested CMOS sensor has a " rolling"
Closure of the non- switched all the pixels at once, but offset by line and time
. This leads when reading a moving image to strong distortions for Astronomy , the
camera is therefore only useful when the telescope the Image holds firmly in place,
which is rather rare in amateur telescopes. Also strong movements by air turbulence
are there less desirable. With its huge field of view and with its high frame rate
with a small field of the sensor would be ideal for the moon and planets (but not
for IR photography), there are also times a version with global shutter come on
the market , one would obtain the CCD camera , a serious competition. The end of
2010 it is to say wait as development continues , especially as new CCD cameras
on the market , the chips come with greatly increased sensitivity in the IR have
interesting because increased luminance images are taken in the infrared at the
planetary photography because of the then less disruptive air turbulence.








Cameras Status 2012 : Of course , there are now several good planetary cameras,
but also the trendsetter TIS has "upgraded" with some delay , and now supplies its
well-known and popular 21s cameras with the 618er CCDs from Sony. Thus, the suitability
of the cameras for planetary imaging has improved again. The new cameras have the
entire area a significantly higher sensitivity to light, at least you realize that
the " blue " End , quite dramatically , the effect in the infrared. This is a great
advantage especially for shooting in rough air , because then the IrRGB method (
see below ) can use to still get enough good pictures. Compared to the old version
, the sensitivity increased in the infrared to three times. We have taken this
as an opportunity , a detailed test of these cameras compared to the previous versions
for "Stars and Space" to write , published in issue 12/2011 with a circumference
of 9 pages. For long-term exposure, the nine 21s cameras are no longer optimal
, because they show a relatively strong noise. The test report from SUW the new
21er cameras was at least temporarily, in moderate quality
an this point are available. It is a PDF, so please be patient for a while
until it is loaded. In accordance with reduced quality is also the first report
in Issue 6/2008
als PDF available.








Conclusion Camera selection : For the planetary photography the DMK21 because of
the high photosensitivity a good monochrome camera, color separations for the toUcam
is quite good despite of data compression. Ideal Cameras for Lunar Photography
(but not füill , the optimal solution would be a B / W image with the DMK21 in combination
with a color separation of DBK21 or toUcam. The lower sharpness of toUcam then
does not matter , because the eye only sharpness of the Luninanzbild looks. An
example of such an image can be found
here , the color comes from a DBK21, the luminance provided a DMK21 in IR
invisible.




Converters and Barlow : On the system of the camera almost always a part of high-resolution
photography a focal lengt h of the optics. Because normal eyepieces are calculated
for parallel beams exit , eyepiece is only with special Projektionsokularen optimally
, or with a very expensive flat field converter as eg is offered by the company
Baader in perfect shape. Normally, however, reaches a good Barlow fine. Definitely
is excellent apochromatic doubly - Barlow Celestron with the can , depending on
the distance to the CCD chip extensions of 2 to produce 3-fold. Even with reverse
insertion of the disassembled base in the beam path ( to come closer to the camera
) it still works without any problems. There are other blameless Barlow on the market
, such as the Televue Powermate , but also inexpensive parts can certainly function
optimally. However, a bad Barlow means the end of all dreams , you should not save
at the wrong place.




Filter: The system of the camera almost always include a filter for high-definition
photography. The use of filters is not a trivial task , first we 'd like to think
that the protection of the CCD chip , a clear glass or an IR cut filter (often in
the form of an IR / UV cut filter ) placed in front of the camera. The filter only
the highest quality are announced here , is self-evident. Almost all of our moon
pictures were taken with such a filter in front of the camera. Under certain conditions
, however, other filters are useful or even mandatory. There are at the end of the
tutorial yet comprehensive information. Who must, however, work with a bad color
corrected refractor or high air turbulence , so can now even read the appropriate
section.




Protection and Cleaning: To protect the CCD from dirt and dust necessarily belongs
actually a filter glass in front of the opening. A speck of dust on the CCD is deadly
for the image, a speck of dust on the filter a few inches in front of it , however,
is entirely unproblematic. Often, you are already using an IR blocking filter ,
which then takes over the protective function , if in doubt do I buy a filter made
of clear glass. Should now but once get dust on the chip, so you remove the with
a clean cotton swab in a dust free environment. Caution: Do not press too hard with
the chopsticks on the extremely thin and fragile window of the CCD ! A successful
control is possible by sticking with the pin a pin hole in the cap of the camera
socket and the camera distance in front of a bright light source holds ( pinhole
camera ). If the dirt firmly , is often sufficient breathing on the chip and prompt
scrubbing of the mist with the cotton swab , very slight bubbles is also allowed
.....











Part 3 : Adaptation of the CCD camera at the moon / planet images




problem : For any photographer raises an important question: " Which aperture
ratio I have to choose to use the resolution of the optics in a particular CCD camera
full" ? At this point I do without specifically on theoretical approaches to the
derivation of a formula for the camera adjustment , because it is difficult to capture
this all influences on the image. Only at the end of the section I give a short
notice for such a procedure and its outcome. As so it goes without theory ? In such
cases, the physicists made ​​a "model" , derives formulas here, and if these formulas
in practice, describe the process well , the model was good. I even thought up two
different models that provide the identical result is both pleasing




1 Model with Dawes criterion: Between two due to the resolution just separable
bright points should still be another point logically on the CCD chip , otherwise
they can not be separated in the image. So there must be such that the two bright
points fall to two pixels , between which there is still a third pixels , so the
image must be distributed over three pixels. From center to center of the two outer
pixels which is a distance of 2x (where x is the width of a pixel ). The focal
length F of the optics must be adjusted so that the predetermined by the resolution
of the smallest dot pitch is accurately depicted on this route 2x. In our optics,
F is calculated from the value specified for 30cm aperture and 550nm light wavelength
0.38 " of the Dawes criterion , as we have already shown in our treatise on the
resolution.




This leads to our optics and camera with the valid for small angle formula sin a
= the equivalent of 2x / F (a resolution in degrees , ie 0.38 / 3600 and x = 0.0000046
pixel pitch of the CCD in meters) to a focal length F in meters F = 2x / (tan a)
= 5.0 m a focal ratio of f17. Generalizing this bill , then a simple " rule of
thumb " results for the aperture ratio of f = 3.6 x, which is valid for any optics
and s / w cameras at 550 nm wavelength (insert x in micrometers ).




2 Model with the Airy disk : We can think like a Airy disk is imaged on a
CCD chip when the image allocates a certain number of pixels. The figure below shows
the schematic for squares of 2 x 2, 3 x 3 and 5 x 5 pixels. In central figure on
2 x 2 pixels , all pixels are uniformly gray, fine resolution is not possible this
way. The 3 x 3 pixels resulting in a figure " crosses " can lead to stage and unclean
representations in long-drawn structures and edges. Only when grid with 5 x 5 pixel
shows a reasonably decent circle-like playback. The mathematical approach for the
three cases gives the equation of the diameter of the Airy disc 2.5 ? f ( For professionals
: That is about twice the half width of the PSF) with the edge length of the square
n × n pixels : 2.5 ? F = nx. Which leads to n = 2, n = 3 and n = 5 f at a wavelength
of 550nm to the below formulas to in the figure , the aperture ratio









As we have seen elsewhere , is also a planar image sets together because of the
superposition principle for waves from the superposition of individual pixels (
Airy discs). For best fit of the camera, it is essential that these minute structures
of an image are also displayed correctly. This leads necessarily to the red frame
in the figure above values ​​for the aperture ratio of f when the camera is to be
optimized. Thus, this approach also leads to the previously derived formula for
arbitrary cameras and lenses : f = 3.6 x. It is for color cameras with a Bayer
mask to set for x , the effective pixel diameter , the 6/2008 about 1.3 times larger
according to our research of edge sharpness in SUW than the geometric value of x.




Conclusion: In order to achieve the theoretical resolution in addition to
good , excellent collimated optics and good seeing is required. We used to think
that the seeing must be perfect. Recent observations show an astonishing way , however,
that even with no means perfect conditions even best resolution can be achieved
with a 30 cm optics. This new finding leads to the following recommendation :

• Optimum adaptation a CCD camera with a pixel grid of x micrometers requires
  in -dimensional objects a focal ratio f , the value f = 3.6 x should not be less
  possible , over disc below this value is sufficient even at light only useful in
  very specific circumstances.
  
  
  
  
  


• optics up to 20cm : owners of small lenses of high optical quality up to
  20cm are sufficiently good conditions find relatively fr equently , and therefore should the above values ​​always possible to comply.
  
  
• optics up to 30cm : Even with lenses up to 30cm sufficiently good conditions are more common than suspected , a valid resolution for optimum opening ratio is also certainly not a mistake, only in really bad conditions or poor optical quality should be deviated from.
  
  
• optics up to 80cm : optics with more than 30cm diameter, the opportunity for sufficient conditions on maybe 10 nights a year drops , who does not want to wait for , can also work with smaller aperture ratio , after motto: A Porsche is still fast even at half engine power. Who works so but should be aware that he gives the best conditions , the added rarely present resolution.
  
  

Working with smaller f - values ​​ The adaptation of a camera can also be theoretically derived via an approach with the PSF. Neglecting this other influences on the imaging (noise , smear , etc. ) , one obtains the formula f = 3x for something smaller aperture ratios. But out of practice , we consider the calculated with our model f values ​​( f = 3.6 x ) will always be meaningful if the optics has a very good quality, and agree to the atmospheric conditions. Due to the required areas at the correct exposure time of less bright edge region of the Airy disc is barely visible, which reduces the diameter , the consideration towards " better resolution " moves , and would require an even higher aperture ratio. Also, to compensate for errors in the electronics of image processing is a certain " oversampling " occasionally the more favorable , the worse the electronics. Consequently, many well-known lunar / planetary photographers work with even larger values ​​, professional equipment like the Hubble telescope do not have to deal with such problems , and therefore also come with a smaller aperture ratio.

Men should also note that the resolution in the considered borderline cases no longer linearly related to the aperture ratio. A doubling of f does not lead to twice the resolution , and who is willing to renounce in favor of more light and image area on slightly resolution , with lower values ​​should work ( but possibly not the values ​​in the blue box below top).

Owner of average optical quality should reduce the f value up to f = 2.2 x especially in poor conditions. Advantages of a smaller f - value are the large field and the short exposure time. However, the exposure time does not play a decisive role as is often assumed in good conditions. The image processing requires at small f - value is a post-magnification by means of interpolation , and is much more critical in terms of jagged edges representation and noise (part of the noise can not be avoided by short exposure , and can be at the occurring at small f - value extremely fine detail no longer separate from the useful signal , see also Plato , the two images above). Who wants to live with these limitations , can easily achieve astonishing results with smaller f - values ​​, which also shows a comparative two pictures that were taken at random with the same look and the same camera at the same time at different locations. However, showing the f10 shot and Nachvergrößerte picture very clearly the limitations of this technique .

Especially in the planetary imaging a certain oversampling to generate clean images is very useful , therefore working virtually none of the known planets photographers at smaller lenses with small f - values, but with values ​​that are even above our recommended values. A good example is the absolute for the size of the optics of only 20cm aperture prime Jupiter consumption by Torsten Hansen. His camera is adjusted properly even at f20 according to our reckoning , the picture was but at f33 , and clearly shows the superiority of larger aperture ratios in good conditions. With a focal ratio of f10 , the moon Ganymede would have been in area 10 times smaller, and could have shown no more defined structures. Even if you had no more fine noise components at f10 from the image can remove.

How accurate are our calculations for adaptation also shows a picture that has been edited for testing a sharpening filter. This image shows many small craters that extend in 1:1 scale only over a distance of four pixels. If one were to take the picture instead of the aperture ratio f20 f10 , this crater would be mapped to a length of 2 pixels, which is of course impossible. Even clearer that is in the number of pixels involved in the f20 are to 4x4 = 16, 2x2 = f10 only a maximum of 4 A crater image with 4 pixels is no longer conceivable as opposed to an image of 16 pixels , and would at best in course of at f10 inevitable image magnification to a generated by interpolation artificial crater-like structure. So who wants to take a very good look at very good conditions optimal images (especially of planets) , so be explicitly warned about the occasionally propagated to small aperture ratios

Part 4: Preparation and implementation of a recording.

Kollo Imation : For high resolution photography an absolutely precise and accurate collimation ( alignment ) of the optics is essential. Since almost all of the mass production, the telescopic mechanism is not executed optimally , there are problems. In a SC, in which the baffle tube (on which the mirror is moved for focusing ) is not quite parallel to the optical axis, the mirror moves in the refocusing of the visual in the photographic position laterally relative to the axis , which inevitably leads to Dekollinmation. To avoid a permanent new collimation , one should therefore attach through appropriate extension tubes the camera so that only a minimal refocusing from the visual position out is necessary. It can not be said often enough: the minimal error in the collimation makes all hopes of a high resolution image to naught !

Extremely important : We initially collimate quite "normal " by centering the intra / extra-focal images. But then comes a very important exercise that you should perform is absolutely correct in any case , if the optics also only has the lowest non-spherical aberrations. Non-spherical aberrations result in air turbulence that Light eruptions not taken place evenly in all directions , but have a preferred direction at the focal collimated image. Suppose , carried the light eruptions caused by air turbulence preferably to the right. Then later in the superposition of the individual frames of the video no round spot emerge , but a extended to the right oval. This line-shaped deformation occurs at each point of the image, and leads to an extremely unattractive appearance of the otherwise very sharp image that many photographers can not be explained. To reduce this extremely annoying error , proceed as follows: the image of a not too bright star on the monitor is observed in very light air turbulence. Upon close as possible focusing ( the already usual collimated telescope ) should caused by the air turbulence "outliers" be distributed at the focal image completely symmetrical on all directions! If the not the case, so you have to correct the collimation so that enters this state as precisely as possible. Otherwise, there is the later -mentioned problems , paradoxically, just when sharp images attract attention particularly unpleasant.

Temperature Compensation: A very significant problem may be a lack of temperature compensation in optics. To test, looks at the highly enlarged , defocused star image , in which rings even the finest currents in the tube , for example, are visible as vertical lines. The generation of high-resolution images is already significantly disturbed by slightest currents, because the finest details are distorted in the pictures. Also in the flow-free optics in a wedge -shaped " cold air lakes " may form on the mirror , which act like a prism and cause nonlinear distortions. How to fix this error (fan in the SC , isolation of the tube ) you can safely search the Internet , we have as yet no experience , and throw heavily loaded videos for now just go away.

The Seeing: In high-resolution photography good conditions are urgent requirements . We provide our tube time to cool down to the outside, but will be established only if the observation with a small optics shows a good chance of seeing. If the object is not high enough in the sky, so we do not quite. Among the problems with seeing that we can add atmospheric refraction ( the envelope of air acts like a prism ) that easily to a range draws the image of the moon apart, what can not be correct in a monochrome shot without filters. Many photographers use therefore ( and this is also true for planetary imaging ) color filter (green , red, IRpass ) to correct the refraction, and the influence of the seeings on the image to reduce ( long wavelengths are not as strongly affected by the atmospheric turbulence , as short wavelengths). However, this increases the exposure time , which is rather undesirable. This quite promising method You should definitely try it , see more details in the addendum to this tutorial. Another important tip : Practice has shown that it is very quiet at dusk often a brief period of air , can be obtained in very good images , timely Beginning may be worth so ! Is the seeing with respect to the high resolution borderline , we nevertheless tried it : More often than you think can edit the videos well, though the air was not very quiet, but the opposite can happen , unfortunately, an amazing way ! A little help in assessing the seeings could look like this : Starks " Wabern " and distortions in the image need not necessarily mean that one has no chance , but it is a prerequisite for a decent result that at least look as the picture sometimes looks quite neat sharp. This is not the case , so there is not no good result , even if otherwise the image appears relatively quiet....

The Mount: As banal as it sounds, it could nevertheless be important : The mount should always be balanced in declination and right ascension so that a small train in one direction is left. If one does not (ie at 100 % balancing ), the slightest wind could have the optics move by the amount of the almost always present in the bearing clearance back and forth , which is certainly not desirable in a very smooth mount. When the moon shots , it makes sense to adjust the tracking speed to the moon up, if the electronics allows. If you record with at planet with the cutting function of the camera only the very narrow range around the planet , it is also useful to compensate for the errors of the screw mount , so if any one use the corresponding function of the electronics! The more accurate the mount is polar aligned , the easier it is recording. Even in our ancient , rickety mount remains a planet for at least 10 minutes exactly in the center , and the hand box must not be used for permanent correction. To succeed also videos that centered later from your computer without any problems and can be averaged.

Focusing : Before recording should be very precise focus , and make the adjustment of focus again and again , you run any risk that all videos are slightly defocused and worthless! When focusing is not limited to the To ensure sharpness , but also for minimal unrest in the image. Often the point of minimum disturbance results in a better criterion for the correct focus as the focus itself If you later uneven sharpness in the images fixed , so you should check the collimation, may see an image plane is not perpendicular to the axis of the camera. The modern Crayford separations can be the cause. We had two of these extracts, in which the moving part was installed slightly askew , and could not be adjusted. To test, clamps the whole extract in a large sliding caliper , or laying on the stationary on a flat surface extract a spirit level. Even the smallest mistakes are fatal in high-resolution photography , and should be left to the part of the dealer to swap !

Setting the software : This is a broad field and can not be exhaustively presented here. The names of the controller and the function also differ depending on the camera software you use. Even with identical software can work different controllers depending on the connected camera, for example, is in the imaging cameras Soruce the case, depending on whether the cameras use Firewire or USB. The delivered to the TIS cameras software IC Capture is but simple and intuitive to use , additional remarks we made in our TIS camera review in SuW 6/2008. Some basic observations that apply more or less for all video cameras , but still seem to be absolutely necessary. Gamma , gain, and exposure time should be in principle possible, always chosen so that the full range of the histogram (with IC Capture hidden or shown ) is used for the region of interest of the image signal , but you should this note:

• exposure time and frame rate (fps ) : This size is the most important of all , and found to be almost 100 % the quality of a video. As long as the chip is not overexposed , the more light the better. If so does not force the seeing at very short exposure , choose a longer exposure time. Very important is still the combination of the number of images per second ( fps) and the exposure time. Since the quality of a video increases with the number of total recycled photons , light is not wasted. This is achieved with correct settings : At 30fps provides you an exposure of 1/30s one to 15fps one takes 1/15s , etc., etc. This rule is strictly observed especially in low-light planet.
  
  
• gain ( Gain) : This setting causes an amplification of the image signal , with virtually all interference signals are increased proportionately with. The position of this controller thus has little effect on the image quality , and mainly serves as a setting of a signal level suitable for further processing of the image, and on the monitor shows a decent picture. The poor signal -to-noise ratio (SNR) at a high gain is not primarily caused by the high gain, but with a too short exposure time. Thus, these must be increased. If we reduce only the gain, so might be reduced visible on the monitor noise due to the lower contrast image , but it remains proportionately fully contained in the image signal. If the gain knob to the far left (small gain) , so the exposure time is too long, the image is blurred by overexposure of the CCD , and bright areas are no longer differentiated.
  
  
• Brightness: This control adds a constant brightness at all gray values ​​of the signal and should be set so that a black surface in the original object on the screen ( almost) looks black
  
  
• color setting: If you are using a color camera , one should adjust the color balance and saturation really sure before recording. A subsequent correction in the context of image processing is not always straightforward and sometimes even impossible.
  
  
• Gamma adjustment : The neutral gamma value is actually called it first but do not hold the camera manufacturer. In the TIS cameras which is especially severe , which is actually identical cameras with USB or Firewire follow completely different industry omen , and therefore have starkly different gamma control ranges with non-comparable scales. In the normal setting (actually gamma = 1, but as I said in almost all applications also completely different, it helps just try ) the 256 gray levels of an 8 -bit image are uniformly distributed over the image signal. This is bad for planetary images because there just some slightly darker areas to be photographed on a very bright disk , which are then possibly distributed only to the brightest 40 of the 256 possible levels , the remaining 216 steps remain completely unused. If, now , a gamma value is smaller than 1, of the 256 possible gray levels for the bright area of the image are more reserved and not wasted on the dark part does not exist . This leads to a on the monitor darkening image but much more details on the bright planetary disk is because now may not be 40 but rather 120 of the 256 available shades of gray are used to display the bright areas of a practical example, if the Firewire camera (not USB ! ) TIS the gamma slider is on schedule recordings on the far left at the stop. Small Gamma values ​​are announced for the Moon Photography off the terminator and at full moon , the terminator lead to small values ​​, of course, " flooding " there also the existing dark areas , which sometimes necessitates a higher gamma value.
  
  
• image section : For planetary imaging , it is useful to include only the planet itself , and to avoid a large black environment. This reduces the size of the video is extremely strong reduced, and also the subsequent image processing often runs faster. This is done using the ROI (Region Of Interest ) function, which provides the software of many cameras.
  
  
• Recording Time: If all settings are finished so still asks how many frames should have the video now. Formulated coarse noise decreases with the square root of the number of recorded pictures. A quadrupling of the number of frames each so bring a doubling of the image quality. An increase from 200 to 800 frames brings Same as a further increase from 800 to 3200 and from 3200 to 12800 frames. So it makes no sense to drive the games too far , especially since the rotation makes for some planets then unpleasantly noticeable. We work with the moon 800-1000 frames and frames planet with 2000-5000 , of which we use later 20-100% depending on the seeing.
  
  

Subsequent gamma correction ? A subsequent correction of the gamma value for the image processing is possible if the gamma control the camera starts only after the AD converter , and the artwork is processed with at least 16 -bit. Since we do not know whether the gamma control already during digitizing plays a role , and also the processing software ( Registax , Giotto , AviStack ) definitely reference points etc. better place in high-contrast images as in " slack " recorded with high gamma videos, we adjust the gamma value during the recording carefully and optimally .

Part 5: Normal image processing

Problem Description: Because this text is intended for beginners , just some basic explanations. Moon images are by air turbulence ( seeing ) almost always blurred and distorted. Sharpness and distortion vary with time very fast, there are also moments of sharp and undistorted image. Unfortunately, this always affects only a small part of the image , the image is divided into small areas that are never shown all together at the same time well. Remedy the processing and evaluation of many images of a video film with the computer in two ways : First, a large number of images is evaluated , it adds the incomplete information of the individual blurred images into an image with high informational content. At the same time reduced the inevitable noise in digital images to the square root of the number of images. I process 900 images so the noise thus decreases by a factor of 30 was only with reduced noise , it is then possible to use an effective filter for image sharpening. Second, the programs can handle many small areas of the images independently , and so counteract the sharpness losses and distortions caused by bad seeing. For this type of evaluation, several programs are available that work with slight variations on the same principle , which is explained in the following tailored to the AviStack program shortly.

Basic sequence of processing : First, the computer analyzes the state of all images ( frames) of the video, and shifts the frames so that they are located as accurately as possible about each other. The thus aligned frames are averaged to form a center image. Then the computer divides the middle image into many small sub-basin, in the middle of any area , a reference point is set. A reference point is a striking structure as possible , which should find the computer later sufficiently secure in all video frames again.

Now, each video frame is divided into regions whose image quality is determined based on specific criteria. The user specifies which quality he calls at least . As a result , the computer searches for the reference point of the previously defined sub-regions in all frames of the videos that you have a sufficient quality, and centered all sub-images based on the reference point for this sub-area. This operation must be carried out for each sub-basin, it is finished for all sub- areas , the centered partial images of all subregions are superimposed , and the assembled the final image. An impression of how the proceeds are the following summary for the processing of a small section of a large moon image :



Software AviStack 1.74 : end of 2008 came with AviStack a new program for processing of the moon and planets videos in use , which also works according to the above scheme. We then looked at after an initial waiting in early 2009 this program in more detail , and created the whole series of new moon pictures from 2009 with this program. There are a number of parameters that determine the flow of the program that we have set in the following manner :

Step 1: First, we put in a good selected video frame two alignment points , which should be as concise and far apart. Ausrichtgebiet 25; Search radius 20; Smoothing factor 1 Then we start the alignment of video frames. If at the end of some particularly pronounced outliers be visible in the result diagram , so we hide these with the controller on the left image, and calculate a new reference image .

2 Step : First, we set the thresholds for any black sky portions on the lunar limb , which should not be covered with reference points. Then, we create the reference points ( the number of Ausrichtgebiete determine ). Smoothing factor of 20 ; Minimum distance 30 (a higher value is not set ); Search radius 5; Correlation surfaces radius 24; Frame extension 96 ( what this is, no idea ). Now we usually store the data for later batch analysis.

3 Step : It was the order of the quality field size, we choose not less than 80 , and start the calculation. As a result , we set the quality threshold to 10-30 percent , and start the alignment of the frames. Then at the end of the superimposition of the frames and storage of the final image. On the sharpening with wavelets we can do without. The alignment of the frames requires longer computation time , which is why it is advisable for the third step, at least for large images of the batch mode using the data stored in the second step data. The final image needs to end yet to be sharpened with a suitable sharpening filter , for this we use the filter by Giotto (see below at " Giotto ").

When processing AviStack provides a set of charts, inform one of the quantitative results of intermediate steps , and thereby facilitating the user to the targeted adjustment of parameters. The following figure shows the result of an initial alignment of all frames of a video. Through the red area , the user can exclude very far shifted frames of the processing , but unfortunately only from above , a little lack of this feature. In the example shown, this function would be at maximum displacements of only 4 pixels , however, unnecessarily , a strong shift should also not be confused with poor quality. The maximum displacement is actually a clue as to who should use search radius at least later (without correction 4, with correction 3). The lower diagram shows that the x - position ( right ascension ) was constantly corrected with the hand box during recording, while this was not necessary with the small drift in declination.



Unlike Registax (described below ) AviStack works are handled properly and absolutely stable without any crashes in the background and batch mode. Intermediate states of processing can be saved at any time , steps can be repeated any time with different parameters. The sequence of the processing is straightforward and more accurate set than in Registax , what rough operator error (as in Registax4 possible) difficult. Compared to our best practices in Registax AviStack expects much longer , and the results were at best marginally better than our Registax method at our pictures. If the notified for version V2 complete batch processing is finished, however, the computation time is not more important , and at an unbeatable AviStack would " tip" for convenient working. Weaknesses , especially in the according to our experience unnecessarily large number of areas in which the image is decomposed automatically . Where in Registax by hand and verifiable set about 15 -25 regions for a 1280x1024 image were necessary and sufficient AviStack generated thousands of tiny areas , as reflected by adjusting the parameters accessible not less than 500 Can press areas. Fortunately, the developers of the program have announced that as of version V2 , the number of reference points can be varied over a wider range. The reference points necessary for processing ( one in each region ) are placed in areas where due to a completely smooth image area not required for such a point structure can be seen. The developer of the program insure but which is present in somewhat larger if necessary correlation surface chosen radius sufficient structure.

If you change to test important parameters of the machining affect the number of sites , the superposition and the quality test , no significant difference in our pictures can be seen in the final result. Also, this suggests that at least for our pictures in the number of Ausrichtgebiete an unnecessary expense is driven , especially RegiStax generated with correct operation similar sharp results with up to 100 times (! ) Less reference points. The developer of the program refer to poor quality video material, many reference points are more important in the processing. In the composition of the image from the individual sub-basins have occasional disturbances in the form of two parallel lines at the seams. Nevertheless, outweigh the benefits of AviStack , so we it already prefer 1.74 Registax4 in the version. While RegiStax fully overlaid by movement when shooting shifted videos, and so a much larger image supplies as corresponds to the format of the camera , this is at AviStack in the tested version unfortunately not possible, show the pictures of AviStack in rough video footage a relatively large blurred edge , because there is no near-edge reference points can be set. This "lack " but should also be resolved with the version V2. With videos over 1GB file size comes AviStack 1.74 for the area set to Registax4 easily clear. With regret , we found that even AviStack relies on the similar wavelet sharpening used in Registax4 , hopefully this results in a better compared to Registax4 solution. I would love to see it when AviStack the unsurpassed good Mexican Hat could offer filter by Giotto ! For the processing of planetary images we have not yet used the program , it remains So initially unclear whether the very good function of Giotto can be exceeded by planetary images.

Software AviStack 2 : This software works much like the previous version, the presentation of the program ( surface ) but was greatly changed. A big advantage is now possible completely automatic operation of the program. We have first experiments with AviStack 2 taken to Jupiter and found that the program is not necessarily very easy to use . The results were not always better than that of a simple to use programs in the planet. In our experiments , there were certain areas of Jupiter with AviStack better results than Giotto or AutoStakkert , particularly in structured enough concrete edge areas. Close to the edge areas in the polar caps were often better represented by Giotto, even if automatically generated at AviStack close to the edge reference points have been removed manually. Particularly with very good video , it did not really matter whether you've worked with Giotto, AutoStakkert or AviStack , the effort in AviStack was greatest. This effort is worth it probably only in bad video , but anyway we do not edit. Although we have experimented extensively with the program and the various parameters , we are still never came to a final conclusion , the results were better from case to case with AviStack times and times with Giotto . Whether AviStack generally now provides for the processing of planetary be nefits justifying such an incorporation into the relatively complicated matter, so as to remain open , the better the video material.

Quite different it looks in the processing of lunar images. Here AviStack 2 is due to the possibility of full automation is an absolute " must" when large amounts of data to be processed , and that justifies the considerable effort to the training in the program. AviStack 2 also working with a small number of reference points now allows ( particularly important for very large images ) , and the inclusion of peripheral areas in an overlap of the video frames caused by shifting during recording. We will report more about the program on occasion.

Software " AutoStakkert ! " : AutoStakkert is far one of the easiest programs to be used for superimposing planetary video in black and white. It also works extremely fast , where other programs require 40 minutes , is AutoStakkert sometimes even in 5 minutes ready. Of course, a substantially lower cost than, for example, is operated AviStack 2 , but the final results Only with optimal settings for all parameters possibly slightly better at AviStack 2, and this optimal setting is very difficult and tedious for inexperienced users to find. When working with unsuitable parameters AviStack 2 works even worse. The results of AutoStakkert are similar to those of Giotto, also a not very complicated -to-use program, but AutoStakkert is even easier to use and faster. For moon shots AviStack is unbeatable, and for color videos of planets is recommended as before Giotto, the people it quickly and easily like should for their S / W video but access to AutoStakkert.

1 Open AVI Flies ( s ) : opens the desired video (or several videos). The box "Write Quality Data " you can usually ungecheckt , the " 1.5 xDrizzle " option brought in our experiments no benefit and also remains ungecheckt. With " COG Threshold " is for setting the brightness threshold to be detected by the on the planet. Only when the planet appears extremely dark , you should choose a value under 20, only when the background is not black, a value greater than 20 is required. Selected with the sliders for image size is the smallest possible cutting, Program works faster.

2 Set Reference Rect : Created by clicking with the mouse on the upper left corner and then click the bottom right corner one is used , a rectangle whose content for superimposing the images. This rectangle must have a contrast as possible of the image wrap. You can also with Saturn mark the left part of the rings , and in a second run, the right part , and together the two images later. If you only want a pass , then you select the Complete ring. At Jupiter , we 've placed all the planets in the rectangle. It then selects " Edge " or " gradient " for the Procedures for auditing. In " gradient " of the contrast in the interior of the selected rectangle is used for the test in "edge" , the edge of the planet in the entire image can be used. Fuzzy edge areas on the planet terminator can be deselected. "Noise Robust " allows the choice of noise filtering in the processing , the stronger the noise , the higher the value. for normal quality is 3 a good attitude. This filtering is not (! ) Applied to the end result , but is only active during processing. The " Invariant to Brightness " is not explained , probably that is a function that makes the quality inspection regardless of fluctuations in brightness of the frames , we have the not tried. With "View Random Frame" , a different frame of video as a reference for further Processing are selected , the better the selected frame , the better the result. Unfortunately, this is a rudimentary function , because you can only move forward through the video.

3 Process AVI: triggers the processing of the (or the ) selected video. The four steps Q / R / O / S to run , the last step is to stack S. Before triggering the processing one selects "Output Images," the desired 16- bit TIF or PNG, and sets in the windows the desired number of stacked frames. Window with 0 are not active. Example: If, in percent of the values ​​50 , 75 and 100 , you get three later evaluations with 50% each , 75% and 100 % of the frames, of course, the best quality frames are used.

the OPTIONS " Normalize Brightness" automatically sets the brightest point in the image to 70 % Brightness, the " convolved images" option creates a sharpened image to each raw image to check how good the raw images. We have these options not used. " RAW stacks " of course must be chosen so that the unprocessed result is stored , and "Save in Folders " makes sense , because then for each of the percentage settings , a separate folder for the result in the original directory of the video is created. "Save ordered BMPs " you should never use , otherwise you will find thousands of images on the hard disk. In the " Prefix" field , you can still specify a prefix for the name of the stored results.

Under " Image Convolution Kernel " can be for the Sharpening kernel used set , which is uninteresting when used for sharpening other filters ( Mexican Hat Giotto ). Under " Reference Frame " you put stack size to about 100 , and uses the supplied Reference kernel ( "Load Reference kernel " and then double click on " ReferenceKernelDefault.afc "). We have the safe side done because we have not quite clear whether the reference frame produced here is used in the stacking of the final results.

4 Castrator : This program is a "Accessories" to AutoStakkert , it allows the preparation of a video in terms of picture and orientation of a planet in the video. The edited video can then be archived to save space , and accelerates the processing at AutoStakkert and in other programs. The operation is simple , the program you should definitely look at times. It reduced Y800 videos of the Imaging Source cameras considerably stores but this may no longer be in 24-bit format , but in black and white and a camera with only 8 bit image depth is probably no big deal.

Software Registax 4: It is essential to take advantage of the multipoint option that works broken down into individual parts, the image When editing moon videos with Registax. Only in this way caused by seeing uneven sharpness can be compensated. Our experiments have shown that many options of Registax extend the processing time greatly , and bring little benefit. We use the simple version with Multipoint "Simple" , and the filter "Gradient " for quality control ( processing area 512 pixels). In the filtering options available in gradient "pixel radius 2 " , we have not changed in the "multi align window" the minimum point distance to 64 is set. We work with about 10 to 25 points in the big DMK41 recording. At the very beginning it is useful to pick out the best possible video image as a comparison image , before they settle on the Multi- Points. When setting the MP , we use the boxes size 128 and small craters in flat Fächen sometimes 64 At the end you should save the selected multi- points, so that they are available when a program error forces a reselection of the video. After alignment , we limit the number of images (limit) to about 200-400 , and lead optimization ("single run " optimizer ) from. Then we select the " stack" , and put the "Feather" option to about 7 pixels ( fuzzy edge when overlaying the parts of the image ) , and the option "expand to maximum image size" , and start to stack. After st acking immediately (!) Save the image ( ! Under no circumstances
should previously be called the mode of the wavelet filtering ! ). Who wants to
even more precisely , can according to the limit with "Limit" or create a new image
comparison of the top 50 single form (create new reference) and easy to sharpen
with the 2nd controller of wavelet. Very rapidly and in a few seconds , can not
hurt it. After optimizing you could still call the stack graph , and individually
or perform for each multipoint an additional quality selection before you start
to stack. This is also quite fast, and can not hurt. The work of Registax is then
done.




Software Registax 5: Registax 4 is a very useful but not very mature , faulty program
. More frequently, nerves malfunction in the operation ( important functions run
but fortunately largely correct) , but you will quickly find ways to circumvent
the error. If serious problems , it is sometimes a complete reboot to avoid the
best solution to further errors. The beginner is confronted by features and parameters
whose meaning and usage does not open up easily or even questionable with a vast
diversity. That this state with the version Registax 5 repaired is to be hoped ,
but to be expected after first impressions hardly. Beginners , we recommend not
to use AviStack. For the processing of the Moon videos and we are currently using
this program. Eventually we will deal later in more detail with Registax 5, and
want to advance due to lack of experience make any definitive statements.




Software Giotto 2.12: This program is the so called " father" the software for processing
of videos to produce images. Giotto therefore does not yet offer the possibility
of an image to eliminate the air turbulence to be divided into sections , which
are then processed separately. It is therefore no longer the first choice for large-scale
images. Giotto works but in the creation of planetary images excellent, and it has
very excellent filter for sharpening. We therefore use Giotto for our planet images
, and to sharpen the Moon recordings obtained with Registax or AviStack.




Software Fitswork 3.99 : This program is a universally usable program with a plethora
of ways. Due to time constraints we have not investigated which of the many interesting
features for moon / planetary photography may be used , but us the Gauss - sharpening
filter of Fitswork have interested. Finally, the optimum sharpening is an important
( perhaps even the most important) process in the creation of high-resolution images.
Examines we have the capability specifically for the Lunar Photography with our
camera DMK41 (see Part 6).









Part 6: sharpening images




Problem Description: Due to the above-described superposition of many individual
images , the noise is very effective largely removed from the result image. Therefore,
it is now possible very effective filter for sharpening applied. In the first still
blurry looking picture is yes by the addition and averaging of many images , the
image information is "hidden " from many pictures, and is just waiting to be made
visible. In principle, this information is therefore provided, with witchcraft has
sharpening nothing to do , even if it is sometimes given the enormous improvement
of the image hard to believe. One can imagine it quite simplistic way: through
the physical properties of the optics ( diffraction , etc. ) and the image of a
point appears by influences of air turbulence is not as sharply defined point ,
but as a small disk , decreasing in brightness from the center to the outside.
This decrease takes place , for example, in the form of a bell-shaped Gaussian curve
, or a similar PSF ( point spread function ). Is the shape of PSF exactly known,
it is possible to use the computer, smeared image points " expected return " to
the original shape , which leads to an enormous improvement in sharpness. Another
approach is the following : An analysis of the frequency spectrum of an image ,
so you will find the finest details in the highest frequencies of the spectrum.
One raises now at these frequencies with a high pass , the finest details are amplified
in the image , the image looks sharper.









Limits the sharpening : All sharpening method naturally have " risks and side effects
." A system based on the PSF method can of course only work correctly if the PSF
is known exactly. But this is in practice never is the case. Also a too intensive
application of sharpening algorithms beyond the limit of 100 % back-calculation
of image blur also causes unpleasant "artifacts" in the image, ie structures that
have to do with the original image content only slightly. The wavelet filters as
mentioned arise in ever greater use increasingly " knitting pattern "-like structures
in smooth surfaces , but also the sharpening without wavelets is not ( as is often
claimed ) largely free of artifacts : When excessive application also arise en masse
fine image structures , which have to do with the original image little to nothing.
" dangerous " Artifacts such as the rings in the dark shadows of the craters could
certainly be removed without serious distortion of the image, but not so subtle
changes in the myriad of excessive filtering. Here, the user is asked to call it
quits at the right moment ! This is in " pretty pictures " merely a matter of taste
, in scientific recordings , however, a very delicate matter.




Sharpening with Photoshop: Since many image editors have this program it is natural
to ask whether this a good sharpening Astro images is possible. You can disable
the filter " Unsharp Mask " USM and the " Gaussian Blur" GW. Latter one has to mention
the same in connection with the sharpening because it is almost always necessary
between the individual applications of the blurring mask. Good results can be achieved
only if the images in 16-bit format and are present in PS and so can be edited (old
versions can only handle 8- bit format, which is not sufficient for proper sharpening
off).




In general, one starts with a strong sharpening with USM , for example pixel radius
1, threshold 0 , 500%. This sharpening you apply twice. The then any visible noise
is mitigated with radius 0.4 GW , then is again a USM with a slightly larger pixel
radius and a higher threshold , for example, radius 1.7 threshold 5 Strength of
350 percent. These values ​​must be depending on the image a completely different
way be selected even number and order of the USM and GW needs to be changed , not
quite a simple matter. An example of a sharpening according to the above scheme,
this Clavius ​​.




Sharpening with Registax4 and AviStack : The wavelet sharpening with these filters
is quite simple, you will find experimentally the optimum for the present picture
material setting of the filter levels relatively quickly. The removal of residual
noise is possible by setting the lower negative image plane with these filters ,
but is only at the expense of image sharpness with fine details. Depending on the
raw material , these filters provide a very good but also less good result , we
have therefore not investigated further. In AviStack 1.74 , the filters are still
experimental, and the new filter by Registax5 we have not tried. So let's wait and
see what results because later......

Sharpening with Giotto. Finished fit images from AviStack are indeed mirrored vertically for some reason , they can be loaded without problems but in Giotto. The product supplied by Registax4 TIF image is for some reason the green channel of a color image , and must be in Photoshop, etc. are converted into a grayscale image , which is then loaded into Giotto and sharpened there. Plain radiographs without Barlow or even Reducer have to edit quite different (this is often even with the wavelet filter RegiStax and AviStack quite well ) , as the high-resolution images. We filter out the high-resolution images , all with Giotto Mexican Hat , and only with the form (!) : Rectangle pattern: square , noise filter : rectangle.

• wavelet sharpening as Registax4 : Relatively fast and uncomplicated, but delivers not for all types of starting material, the best sharpness. The presentation of large-scale contrasts is very balanced and well done. Artifacts occur with conservative application to only a small extent.
  
  
• Gaussian sharpening with Fitswork 3.99 : Not quite straightforward especially because of the after-treatment may additionally required at critical imagery. Sharpening of fine details is possible with Fitswork best , in contrast sharpening of large-scale structures is the Result is not always optimal. We use this filter occasionally , when it comes to maximum sharpness. Artifacts occur especially when the focus is maxed out.
  
  
• Mexican Hat Giotto : This sharpening filter does not work with the absolute maximum sharpening the finest details , the sharpness is still to be assessed as very good, and the image impression is the easy processing even worse picture material actually getting of each feasible optimal compromise. We use Giotto for all of our "normal" Moon images. Artifacts occur depending on the quality of the starting material on the normal scale.
  
  

Part 7: image enhancement.

Professional Noise Filter : In worse conditions , there are some very effective way to improve the image, for example, the use of a professional smoke filter. Such a program is safe because it does not create artifacts in the worst case, but are present in the image structures " ironed out " , which is obviously not the intention. It allows , for example, the complete removal of excessive noise in black areas with simultaneous removal of minimal residual noise in gray areas , or the removal of tiny artifacts that may have arisen in too much sharpening. This all takes place practically handled properly without loss of important details of the image. We came occasionally an older version of NeatImage Pro + program to use. The matter is not so easy to get to , but worth it in any case ! Here's just a small note to the basic sequence. The work is done in two steps:

• First, the program examines the noise in the image , and created a device noise profiles. With a sort of " equalizer " can then hand still changes to be made if, for example in dark areas , the noise will be greatly reduced especially , pushes turn the corresponding control further up.
  
  
• Then one makes the Noise Filter Settings , there pushes you the controller for " Mid" and " Low" to zero , and, with the controls " High" ( for the small details ) and "Y" the desired effect one ( control of a drawn rectangle with the mouse in the picture). The upper controller " High" represents a are considered up to what size details as noise , the controller " Y" indicate how much noise should be reduced .
  
  

One more note on the application of professional noise filters , which also applies to the planets Photography: Who has foreseen from the outset the use of such a program should possibly choose when recording a slightly larger aperture ratio , as in our formula from SuW 6/ 08 for optimal adjustment of the camera is required. Thus, although increases because of the lower light intensity the noise , but which can be better then separated with a professional noise filter from the coarser image signal. For cameras with already low light sensitivity of these methods is recommended, however, by no means ! Even with the sharpening of the image can be a little more aggressive procedures when touched up later with a professional noise filter.

Other image enhancement : The ultimate emergency brake is retouching. For scientific recordings is only allowed if clearly (! ) Recognized artifacts or errors can be removed from the image without distorting the image content , which make professional image editor with NASA photographs occasionally. For us, the barriers are not so high : Before you throw away a troublesome created image , you could , for example, edit an inserted RegiStax improper connection of two MP - areas with stamp or smudge , or black areas in the craters remove overshoots the sharpening. But should we retouch only if a structure clearly (!) Is recognized as an artifact , but it is better to pass up completely , and edit only the best videos. By creating selections ( if necessary with soft edges ), certain areas of the image can be selectively processed , there is a plethora of features for an image to "improve" , it is impossible to list them all here , but you should use very carefully. Important is the following rule:

• The better the starting material is the less must be " sweetened" be. You should therefore only use ordinary videos and also waive the application of a noise filter as possible. All other treatments are anyway be avoided if possible.
  
  

More latitude : When moon pictures very often there is a case that the video camera to the extreme difference in brightness such as between dark areas and bright crater rims can not bridge. This leads with the correct exposure of the dark regions to a blatant overexposure to the often small bright areas. You do not want to live with it , so you have to avoid with gamma and exposure time selection overexposure . However, this leads to an exposure of the dark areas that can later hardly compensate for image processing still. A stopgap measure is a selection with a soft edge around the small bright areas to exclude the latter from the Levels of the dark areas , as we have done it occasionally. The only proper solution would be the inclusion of two differently exposed videos, the overexposed parts ( crater etc.) can then be replaced in the short- exposed image by properly exposed selections. A process that is in deep sky shots of the day.



Part 8: Summary of the results.

As mid-2007 came up with the idea to obtain a decent optics and a good camera , was not yet clear how difficult that would be. Ultimately, this led to our camera test in SuW Issue 6/2008 , and to a very respectable 12 " ACF optics, Meade. Lots of experience in the creation of high-resolution images we did n ot , and there was a learning process for us , the result of which is reflected in this tutorial from 2009. Important for the success of good images , therefore the following points :

seeing - camera - image processing - Optics

The role of the seeings : Let's go from the premise that even in Central Europe occasionally sufficiently good conditions for shots are given with a very high resolution , so this reduces to a point sufficient patient waiting for the right time. The course lasts the longer , the greater the diameter of the optics, we remember : The maximum achievable resolution is directly proportional to the opening.

The role of the Camera: The currently available CCD cameras (but not CMOS) are in terms of sensitivity to light and noise , although at the limit of what one would wish ideally ( in the planetary imaging it is with the exposure time often already badly almost ) , but it's still not a problem at an acceptable cost to acquire a very neat camera , and adjust as described by us on the optics. In planetary images with monochrome cameras still follows the consideration for the correct filter selection to achieve optimal results , a few comments on this below. If you want to create a IRGB or IrRGB with Ir - pass filters , possibly should use a camera with a Exview CCD, these chips have their maximum sensitivity in IR region. An insurmountable problem arises in any case in the choice of a suitable camera.

The role of image processing : Who our little " tutorial " has followed can find easily that the whole affair , at least for high-resolution photos and Moon in planetary images with color cameras actually holds no secrets that we can not master every other beginners with a little practice :

• is added, with AviStack or Registax Multipoint in the simplest settings
  
  
• It sharpens with Mexican Hat Giotto
  
  
• It filters optionally with NeatImage
  
  
• It makes a slight post-processing with Photoshop
  
  

Until the finished image the moon then takes only the computation time of the computer plus a few minutes. In planetary images with color cameras it's way faster , it does the same thing , but used for adding Giotto without any quality selection in the option " centering on bright disk ". Also, to compensate for the atmospheric refraction ( shifting the color channels ) is Giotto well suited. After a short practice time is obtained quite amazing images that have a quite " upscale " level. This is also true for planetary imaging with color cameras , which also provides beautiful results without any closely guarded knowledge as Mars and Saturn . The images shown arisen in a few minutes with the above- mentioned simple steps of image processing as it can make every beginner after a short exercise. If you want to use as a beginner in the planetary imaging filter , should a monochrome luminance image using eg make a red filter or a 685nm IR pass filter , and then combined with the color information of a color camera. This provides a relatively simple manner without the use of a IrRGB RGB filters. The technique for this purpose we have investigated and presented in an addendum to this tutorial below. Jupiter was photographed by us with this technique. The secret of good images is therefore definitely not in a particularly sophisticated "secret" method of image processing , it also works without consuming process x times with complicated processing steps! So we hold fast what should motivate the novice :

• The Secret good pictures at the Moon Photography and planetary imaging with a color camera does not lie in special professional skills in image processing , every beginner can after a short exercise too much respectable results come.
  
  

Professionals will still be out to get in the picture , having all sorts of tricks a little more of her pictures , but the difference is very limited. For planetary imaging with filters , for example, in LRGB , IrRGB , RGB method , this statement applies only limited experience here is very useful , but unfortunately not always passed in the desired dimensions. An example of this is the homepage of Damian Peach , which provides very excellent lunar and planetary images. Although he works with an unnecessarily large aperture ratio , he brings the C14 to the exact we calculated limit of resolution. A salient feature of his recordings is the absence of the otherwise often still existing " gray haze " in the planetary images . Information about its type of image processing you are looking for but in vain on his website , but there is a fee , a DVD that will then probably also does not help much. We find that a pity !

The role of optics: So if now the image editing (apart from the LRGB method, etc. ) requires no special knowledge of why some people deliver outstanding images , while others take it to mean maximum initial successes ? The answer is easier than you think : What is a video that is a simple image processing , even for less experienced cyclists to access, which does not show a video that brings not forth even the best professional. The secret lies in the chosen good optics which all Amatuer professionals have somehow procured. Videos taken with such lenses under good conditions, even any less experienced editors will help make excellent pictures. From the correctness of this statement anyone could convince himself, who has the moon video used to be offered by us on a CD once chased briefly by Registax and Giotto. So we conclude:

• key to success : You can twist and turn as you like, the key to success is and remains a qualitatively superior optics , as every professional has somehow procured. Unfortunately, the required quality have on the market optics to a very large percentage not.
  
  

Here's some good advice , but in terms of quality , there are some approaches:

• One can find a dealer who delivers a surcharge selected optics
  
  
• You can assure surcharge best quality and Returns Policy
  
  
• You buy " known good optics " , such as the C9 , 25, and hopes to be lucky
  
  
• You buy only expensive optics and a binding protocol, eg OMC 300
  
  
• It checks the optics from the merchants themselves on artificial star
  
  
• It sends the opt ics on their own account a reliable (!) , dealer independent auditor
  
  
• you buy an optics (also used) , one has tested itself (as in Tutorial 1 )
  
  
• Man buys and sells until it finally "fits" ( can take many years )
  
  

We examine the optics to dealers the artificial star , the look is in order here , as is a return on the subsequent test , as described in Tutorial 1 , rather unlikely. In this way, we have also come to our rather good Meade 12 " ACF optics. Also helpful is the following: Some traders and companies to promote their products to the skies : Professional-looking , quality of professional observatories, test protocols from questionable sources , manufacturers indicate in the advertising images of perfect star tests that never such a look in practice will deliver. Such advertising claims such as the Internet can be copied , and later submit , if the product does not meet the requirements of the advertising. Exchange or return should then be no problem. But it is important , the dealer clearly to say before buying that the optics for high-resolution photograph is to be used , a haggling at the cheapest price you should not start , but rather provide for the audit work voluntarily a significant cost....

Optics Size: As we have seen in the tutorial , solely the diameter of the optics determines the maximum achievable resolution. Even a slight increase in the diameter of 2 " has enormous influence on the image quality. Here is a brief consideration : In the amateur level have at least almost all lenses from the mass production , a more or less strong spherical aberration ( SA) or zones. This problem often grows with increasing size of the optics. This now causes always falls exactly only the light from a particular part of the mirror surface in the focus, the remaining light is slightly defocused. A defocused image but extremely strong showing any air turbulence ( some astronomers therefore not focus on maximum sharpness , but on minimal movement in the picture). Especially a great look with SA or zones may therefore show a steady image even at relatively good seeing before. Therefore, still persists the idea that the seeing with increasing optical diameter always causes more problems. In fact optics are slightly less seeinganfällig up to 15cm opening , with diameters well above it is also true that in the amateur field , even in poor conditions never the smaller optics, the best photographic result. 10 " So bring in bad seeing no advantage over 20" , play with better werdendem Seeing then the 20 " always merciless their advantages ! So we conclude:

• For diameters between about 15 - 60cm will provide the best photographic results even in poor conditions never the smaller optics. In amateur so regardless of seeing the rule " the bigger, the better " , the problem is solved if you have deep enough pockets.
  
  

To get the most moon or planet - with the desired resolution ascending - unfortunately, so good conditions required , as they are on an average observation space in very few days a year. So you have to make an attempt , which provides only a very few cases, the erwünsche result to every clear night. This is for " average consumer " as that need us their rickety hand - mount always rebuild , a disaster! Professionals often have a cabin with fixed telescope on a regular mount, and can sometimes just trial " see through " if the seeing fit. If does not work you to the cottage , convenient as it gets then. Despite all these obstacles, but should not be discouraged you look. We first calculated as the tutorial described quite realistically , what is the maximum possible with the opening at their own resolution , and then rejoices when it has succeeded in this specification to get close......

Our work is thus initially completed. What we can contribute from our limited experience, we have given here without reservation protocol. Some interesting questions are still left open , and can be possibly later answered!



Part 9: Supplement on the topic filter and infrared photography.

In the photograph of the moon and planets are constantly struggling with problems that worsen the result of the efforts. Three of these problems can get under certain conditions by the use of filters at least partially into the handle :

• caused by atmospheric refraction blur
  
  
• caused by atmospheric turbulence blur
  
  
• Poor color correction optics
  
  

Who now a recipe for using filters expects will be disappointed. When using filters, so many factors play a role , it is essential to familiarize yourself with the effects of the filter familiar. Only in this way can then in each individual case to weigh the benefits against the drawbacks , and come to a suitable selection.

What is atmospheric refraction ? Light of low-lying objects occurs at a shallow angle into the atmosphere , and is therefore broken as by a prism. In the visible light region the refractive increases significantly with increasing wavelength ( dispersion). Thus, blue light is deflected more , and thus appears to an observer in the rear extension of the light path coming from above, red light is less broken , and therefore, seems to come from below. The image of a planet thus consists of several slightly shifted from each other color images , of which the blue image at the top , and the red image is at the bottom (indicated by a blue color fringe top , and a red color fringe bottom, see Jupiter top left ).

How to take atmospheric refraction ? For color images, which is pretty easy. It decomposes the image in the RGB - channels (red, green, blue) , and blue shifts downward , and red upward until all three channels are placed exactly one above the other , see < a href = " http://www.gym- vaterstetten.de / Freezer units / astro / gallery / images / Filteranwendung.htm "> 2 Top left image . When s / w - images a subsequent correction is not possible , you have to during the recording process using a color filter (IR, R, G, suitable ) so that a sharp image reaches the camera. Which filters to choose depends on the air turbulence from , and is discussed further below.

What is air turbulence ? While passing through the atmosphere, the light regions of different density happens. these can often be very small ( turbulence cells). In this case, the light is refracted in rapid changes in all directions , and can produce a sharp image in the camera. Especially bad is when the light path in the atmosphere is very long, so in low-lying objects , the effects on the figure are then catastrophic, photography is impossible. Since in our part of " seeing " rarely really good, of course, makes you start thinking about whether there's a workaround .

How to eliminate the effects of air turbulence ? Subsequent elimination of blurring caused by atmospheric turbulence is impossible (except for the usual sharpening sharpening filters and the decomposition of the image into sub-areas in image processing). Now it is so that the refraction ( refraction) and thus also associated splitting into colors ( dispersion) decreases with increasing wavelength. Red , or even infra-red light is substantially less disturbed than green or blue light. When s / w - images simply uses infrared - pass filter in front of the camera for the picture to improve . This is easily possible because the modern cameras (especially with ExView - CCD) just in the infrared have a very high sensitivity. The improvement of the imaging is considerable, see Mondbilder. Ir - pass filter can easily worsen the differentiation of gray levels of the moon , however. In color it is a bit more complicated. Here, you do use of the fact that the eye does not according to the sharpness of the color ( chrominance) , but after the cut-off percentage (luminance) of an image. So you made ​​with the Ir - filter a useful s / w - image , and sets the color of a useless "normal" A color image. The result is a quite useful Ir - RGB, see Jupiter above right . In less severe air turbulence is also an R - RGB or even G- RGB can improve the image quality , which filter the exposure time not so extreme as to extend a Ir filter. Only in this way we could despite miserable conditions in 2008 still quite nice shots of Jupiter make as Ir - RGB, and not quite as bad conditions in 2009 a picture of Saturn as R- RGB. Another Ir - RGB Jupiter was in slightly better conditions in 2011.

What is poor color correction ? With astronomical telescopes can not be collected without any problems in a single focus all wavelengths ( colors). This results in color fringing to the images and reduces the contrast significantly , because more is only one color in focus. Particularly critical is the case of lenses because glasses breaking the various wavelengths differently. Even extremely expensive Apochromaten still have residual color errors , if only in the invisible IR or UV ( but the " seen" by a camera as opposed to the eye very well be ). Mirrors have no problem with color errors , they are combined with glass (like a SC) shall occur low color error here on. The upstream at the cameras Barlows or converter also cause (often little ) color error.

How to eliminate the effects of bad color correction ? For color cameras elimination of blurring due to chromatic aberration of the optical components is unfortunately impossible. Partial remedy only managed to better optics with a custom lens correction which stands for some refractors available , or special filter to reduce the color errors without the image too much to color (eg Semi -Apo filter, Fringe - Killer) . For black -and-white cameras the problem is different. Here you can use the color filters not let the appearance of the incorrect focused colors. If , for example, an expensive APO in the visible light without censure, but is IR and / or UV do not correctly , you use an IR or IR / UV cut filter. Is a cheap refractor well corrected only for green, so you use a green filter. This can often lead to extremely good pictures , but unfortunately no color images. But there is a trick on how to do go ordinary color images : you make three images with R , G, B filters , and sets these three b / w Images are stitched together in the computer to an RGB color image , a somewhat complicated method , but in this case the only solution.

So All the best ? - Sorry, not quite ! Filters can in poor conditions while significantly increasing the image quality, but unfortunately they extend the exposure time and frequency proportional decrease the resolution of the telescope example Ir - Filter : Up to 30 % less resolution at four times the exposure time. One should choose because of the resolution possible, a filter for a short wavelength actually so . However, since just the short-wavelength blue is strongly scattered in the atmosphere , the choice would rather fall on green. If you want to compensate for atmospheric turbulence , one is forced to the use of red filter or even Ir filter. We have an IR filter 685nm used by the resolution produces a reasonable compromise yet. There are filters that still lie further in the IR , but beware : such filters extend the exposure time enormously , because the sensitivity of the CCD camera often decreases again in this area. Who has a usable for IR optics (and these are , for example, practically all of the mirrors ) , and a sufficiently sensitive camera , which was in bad seeing necessarily even try shooting with IR pass filter. In very good seeing conditions , the filter will only bring benefits if the object at a B / W shot is very low ( atmospheric refraction ) , under good seeing and Erected object , the filter is unnecessary without filters are then (with a few exceptions ) produces the best results.



(+ ) means of different effects in each case. O means : no significant effect
.


Exceptions to the rule ? Under certain circumstances, one should use filters even at best observing conditions. Occasionally can be brought about enhanced details of specific color on the planet through the use of an appropriate filter , one of many Examples is the representation of clouds on Venus with a UV - pass filter or the photograph of Mars with an IR filter.

• Conclusion: filter can dramatically improve under certain conditions, the image quality at planet and moon. But it is to operate at optimum atmospheric conditions and with very good optics at best, and apply filters in the interest of optimum results possible only where it is necessary to emphasize certain details necessarily ....

---

Источник: http://www.gym-vaterstetten.de/faecher/astro/Fotografie/MondfotografieTutorial.htm гуглоперевод с немецкого