The flat field box
A device able to provide an accurate flat field for calibration of digital images.
by Lorenzo Comolli

Dust and vignetting are problems causing defects in every CCD or DSLR astronomical image. While dust can be removed from the digital sensor surface or from the near-focus lenses, vignetting cannot be phisically removed except changing telescope. To overcome the first problem I've wrote a small tutorial about DSLR sensor cleaning. But even after an accurate cleaning, dust can come back in small quantities, and it's a problem. And, moreover, other problems, such as non-uniform pixel sensitivities, can be present. The solution is to use a calibration frame, named "flat field", to correct each individual image, after the dark subtraction, but before every other processing.
A flat field contains only the information of how much light is coming to each pixel. Let's see an example: point the telescope to an uniform background and take an image, a pixel may receive 100 photons and another one only 50; the same ratio will be respected also during the night sky imaging, and so, to get an uniform reply from both pixels, the signal from the second one must be multiplied by 2 (=100/50=reference/pixel to be corrected). Usually a flat field image is "normalized", i.e. divided by it's average value. So the mean value of the normalized flat will be one. To get the coefficient to correct an image, each pixel from the deep-sky image must be divided by each pixel from the flat field image. This computations are automatically performed by nearly all the astronomical processing softwares.

The quest for the uniform source
A fundamental step to get a good flat field image, is to point the telescope to a bright source with an extremely uniform field. Thise sources are quite hard to find to the necessary uniformity, and that's why the construction of a flat field source is important. Many sources can be used:

twilight sky
  • easy to get
  • not perfecty uniform
  • stars in the background (necessary a median combine, after normalization of each image)
  • continouously changing brightness: a few tens of minutes in the twilight can be used with the proper brightness
  • being necessary to have the camera at focus, this method cannot be appied during the evening twilight because normally focus is reached when the sky is quite dark
  • if the morning sky is clouded, no flats can be obtained.
dome flat
  • can be remotized
  • can be used always
  • constant brightness
  • a dome is necessary
  • internal white painting or white painted flat surface
  • uniformity depending on the illumination
flat white panel
  • easy to build
  • cheap
  • can be used always
  • constant brightness
  • not perfectly uniform depending on the position of the light source
flat field box
  • very uniform field
  • can be used always
  • cheap
  • constant brightness
  • quite bulky
luminescent panel
  • very uniform field
  • very compact
  • can be used always
  • constant brightness
  • quite expensive (about 1€ or 1$ per millimeter of aperture)
LCD monitor
  • easy to find (a notebook)
  • can be used always
  • constant brightness
  • maybe too bright
  • uniformity depending on the model
white T-shirt in front of the telescope
  • easy to build
  • cheap
  • can be used always
  • not perfectly uniform
  • direct transmitted light contamination

My flat field box
Considering that I've tested many of the above methods, and that I'd like to have a very uniform flat field, I've found that the best compromise between quality and cost is a flat field box. Many online projects are available, and I've taken the best from each, to build my own with available supplies from local stores.
Key points are:
Here are some images that illustrate the construction steps.

The mechanical and electrical schemes for my 350mm front panel flat box.

The electrical parts used, from the bottom left counterclockwise: 9 leds, 100 ohm resistors, a 1 kohm wire potentiometer, a female-male connector and standard red/black connectors.

Detail of the 9
flat leds with diffused light and 5000 K color temperature

A white polywave panel with the led positions. An advantage of this material is that can be easily perforated by the pins of the leds.

The white panel with the 9 leds and the black panel with the potentiometer and power supply connector.

The backside of the above panels.

The black polywave box must be painted white on the internal side. On the left, one of the two opal plexiglass panels. To fix the many parts, I've used the hot glue. This makes a very rigid connections that was able to resist even after a falling off of the finished box from a heigth of 1,5 m...

Backside of the flat box, with the back black panel removed. The fixing of this is obtained with velcro strips on the four corners.

The flat box from the front, with only one plexiglass panel, in the center. Note that the uniformity is quite good even with only one panel, thanks to the flat leds.

Back of the finished flat box, with the potentiometer and power supply connector

Front of the flat box, with a round mounting flange for my telescope.

Image of the panel uniform brightness, in the dark.

An example of a flat field frame obtained with my flat box. This is the average of 50 frames of 120 s, using a 6 nm H-alpha filter. A long exposure was used to overcome the problem of the shutter vignetting with short exposure times. Note the dust on the sensor surface (black points and gray areas at bottom), and on front lens (doughnuts and filaments). Moreover strong vignetting is evident in the upper right corner, and -smoother- on the whole area, centered. And also horizontal bands due to the CCD response are visible.

A good flat field calibration frame is a must to the demanding astro-photographer. A flat box is a very good source of uniform light. And it's cheap, compared to other solutions: the total price for the mechanical and electrical parts are about 50€, for a 350mm panel. On the other side, the dimensions can be a problem for the travelling imager. But mine is used in my backyard observatory, and so dimensions are definitely not a problem.

Some other flat field boxes:

Thanks to Olivier Gallati, that let me test his 200mm flat box at the end of a wonderful night of imaging.

For any comment, send me an e-mail:

HTML Editing and Publishing by Lorenzo Comolli. Email me at
Back to Main Page