How to image extreme lunar crescents

Martin Elsässer 17.1.2008

After the success with the extreme-lunar-crescent contest of the Interstellarum magazine, i have received requests from several people asking for more details about the techniques used. This text has some more information.

Basic ideas for imaging extreme crescents

  • Capture the crescent during the day, to have more time then during the brief periods of twilight visibility
  • Image in infrared light, to maximize contrast
  • Use baffles in front of the scope, to prevent stray light in the optics

  • BE EXTREMELY CAREFULL WHEN IMAGING OR OBSERVING NEAR THE SUN DURING DAYLIGHT. The concentrated light of the sun can and will destroy your eyes and equipment, if you make a mistake. Beware! Team up with an experienced solar observer, if you do not yet have the required expertise.

    Equipment used

    1. A small telescope with good optical quality for infrared light such as a small refractor. A focal length of about 1000mm and an aperture of 50mm has worked well. The focal length must not be to short, so that the width of the crescent spans more then one pixel.
    2. A baffle in front of the telescope, to prevent direct sunlight from entering the optics. Due to the small small angular distance between sun and moon, the baffle can get quite long. For extreme crescents even the illuminated edges of the baffle itself can be a problem. (A second baffle in front of the first (which must not be visible from the aperture) might solve that, but will increase the length by a factor of 4, growing to > 3m. Not so handy.
    3. A infrared-pass filter, to reject all visible light and thus increase contrast between the moon and the blue sky. These can be purchased for infrared photography and are not expensive.
    4. A high quality, cooled CCD-camera which captures 16bit images and is sensitive to the infrared light transmitted by the filter. I used a Sigma 6303. Good cameras can be quite expensive. Standard DSLRs are probably not suitable, due to the inbuilt ir-rejection filters and insufficient SNR. Digital cameras or video cameras are even worse in that regard.
    5. A astronomical telescope mount with goto-capability to pinpoint and track the invisible crescent with the scope.
    6. A solar filter to put in front of the scope for focussing the optics and aligning the mount on the sun. Remove when taking the actual images of the crescent.



    1. Test your equipment and procedures on a non-extreme crescent. Test you baffles, filters, exposure-times by slewing the mount to the expected angular distance for the actual observation.
    2. Get your equipment to a high-altitude location to get as little water-vapor and dust as possible. The sky should be really blue. This is vital for the more extreme crescents!
    3. Align your mount properly, so that it accurately tracks the stars. Align the mount on the sun during the day, using the solar filter. (Cover all optics not equiped with a filter.) Use the camera image to center and align the mount. Move the mount from the center of the sun to the center of the visible edge of the crescent. Have the mount track the moon or repeat the procedure of aligning-on-the-sun and moving-to-the-crescent repeatedly.
    4. Align the camera orientation with the celestial axis, to allow easier identification of specific features of the crescent: eg: North=up, West=right
    5. Focus the scope with the camera on a planet or bright star, with all the filters you will use for the crescent. The goto-mount will allow you to find such a star. You can also focus on the sun, but this might be more difficult.


    1. The exposures of the camera should be long enough to properly use the dynamic range of the camera. The sky brightness should be around 40000 for 16bit imaging system. There will be no shortage of light, you will probably need additional filters to decrease the brightness to the desired levels.
    2. Capture whole series of images, as the very faintest crescents are not visible in a single image. When these images a processed identically and shown in rapid series, the crescent can become visible through the noise.
    3. Capture darks and flats, so as to simplify the image processing

    Results and more information

    Check out the images and videos on the following pages:
    Image processing of lunar crescents
    June 2007 lunar crescent
    May 2007 lunar crescent, german

    back to Mondatlas main-page