Making of a lunar mosaïc

The purpose of this page is only to show how I realize a lunar mosaïc.Many other technics are available, and if you think that you know a useful trick, do not hesitate to let me know it.
On the one hand, I'm always looking for new tricks, and on the other hand I will make known them to the community through this page,

the shot

This first step is obviously the essential part of the job. If the shot is spoiled, at best you will get a bad quality picture, at worst you will have unexploitable pieces of image that you will be unable to gather together.

Some obvious points that are always worth to remind :

Shot the pictures when the Moon is high in the sky, in order to minimize the effects of atmospheric turbulence. This is not always possible, specially at the beginning and the end of the Lunar month.
If you have a reflector type telescope, remind to collimate it before the shot. This stage is often neglected, and this can change a beautiful picture into an only passable one.
Try to take the pictures during a steady night, with a low turbulence... But this is a parameter which is unfortunately out of our control.
At last, adjust the focus as carefully as possible. Digital image processing can do miracles, but it can't repair a bad focusing. To focus the webcam, I use a Borg helicoïdal focuser (non-rotatif of course) which is light and precise. The focusing mechanism of the C8 is in fact not really very precise and induces shifting.

Now, we must take a great care in setting the orientation of the webcam. The sides of the image must be parallels to the movements of the telescope. By this mean, we will be sure to obtain a total covering by moving the scope along only one axis.
If we don't take this precaution, we will obtain a "stairs step" covering, and unless to be lucky, we will get holes in the final picture. And a picture with holes is a bloody awful picture !

Next, we need to set the exposure time and amplification of the webcam. We must not let them on automatic, but set them so that no area of the Moon will be over-exposed. In the same way, it will be interesting to let the white balance on a fixed value. All this is necessary to minimize the light differences between the individual pictures.
For a full Lunar picture, I tune the webcam on the most brilliant areas of the Moon.

During a growing Moon, the surroundings of Fraunhofer crater (after the fourth day in the Lunar month) are the most luminous parts. During the second half of the Lunar month, you will have to set the webcam in the vicinity of Tycho and/or Byrgius.

For the shot itself, I use QcFocus. I take 20 seconds films, at the rate of 15 pictures/second. Each film is about 130 Mo, for 300 pictures.
You must not hesitate to take a large overlapping between two successive pictures. This will be of a great help for aligning the pictures later, and of course, you minimize the risk of discovering a hole in the final picture.

the treatment

To process the films, I use Registax. This freeware is a good software, even if its ergonomy is surely not the best one (this explanation uses Registax v2.1, I haven't yet tried the v3).

To begin, I choose a particular picture of the film, which will be used as a reference by the software. Unlike some peoples, I noticed that the choice of this picture is important for the final result. That's why I try to choose one of the best pictures of the film.

For the choice of the alignment box - the square in the upper right corner, I only aim at a noticeable detail, the processing is sensitive enough to correctly do the job.

The second stage deals with the alignment of the pictures, and the choice of the right parameters to do so.

Ideally, the alignment filter (or FFT filter) must be located at the end of the curve, and the quality filter must be centered on the elbow of the curve, with a quality factor between 0,11 et 0,16.

The optimization features used here give usually good results for the Moon and the planets too.
Be careful with the tracking : don't tick the boxes 'Predict track' and 'Misalign Warning'. Not only this is of no interest, but sometimes Registax can crash.
And now, we only have to click on the button 'Align'.

The third stage consists of stacking the pictures of the film.
In order to do so, we choose the pictures that we want keep. Usually, I keep between 50 and 60 pictures for each film.

The vertical cursor defines a quality level (fullness of the spectrum) of the pictures, and the horizontal cursor limits the differences between the pictures. The point of the curve which is on the bottom axis stands for the initial picture, that we choosed at the first stage, and which is used as a reference.
The positioning of the two cursors is simple enough when the curve is similar to the one which is shown here. We just have to lop the curve in order to discard the pictures with too much difference, and we adjust the quality to keep the right number of pictures. If the curve is less perfect (a quickly falling quality for example) it may be necessary to try some different adjustments to find the one with the best results.
Of course, the higher is the quality of the individual pictures, the better will be the final picture.
Possibly, we can check the 'Use Image Quality' box, but the differences are often small.

Last stage : we apply a wavelet filter to obtain the final picture.
The values which are used here are quite representative of a lunar picture : about 15 on the second layer and 6 or 7 on the third one.

You can always try with other values, especially if the quality of the pictures is good enough, a small value on the fourth layer can be interesting for some areas. On the other hand, it may be necessary to slightly lower the values near the terminator to avoid artefacts.

The assembling

When all the pictures have been prepared, we must paste them together. To do so, a software like Photoshop (or PaintShop Pro) is really useful. We just have to paste each picture as a new layer and to settle it in relation to the next ones.


By using the mode 'différence', it is easy to settle a picture in relation to the already fixed ones. On the left picture, there still remains a slight gap, but on the right one, the new picture has been perfectly fit in.

Using a blurred edges eraser will cleanly suppress the edges of the picture, and will make the joins disappear.

Here is the image, once all the individual pictures have been pasted.
In spite of all the precautions, there always is some level differences between the pictures (if someone could explain me why ?).
We must now adjust the individual levels of each picture to get a uniform luminosity on the whole picture.

The final result is visible here. The picture has been turned over to appear in the "right" direction, at least the one that we are used to see the Moon !