There are some astrophotographers who have perfect dust free cameras that have identical sensitivity on every pixel. They have clinically clean filters fresh from a computer chip manufactures clean room. Their telescopes are high value apochromatic refractors with such large illuminated fields that you could hold a night-time sporting evening in the eyepiece holder.

These fortunate astronomers probably don't need to do flat fields.

All other astronomers (Which includes you and me) do need to take flats.

As you progress in astrophotography, you will find yourself putting more and more effort into each image you produce. It isn't long before you are running up long integrations lasting several hours. It never fails to surprise me how many astrophotographers put all this effort at risk because they can't be bothered spending ten minutes at the start or end of the evening taking the required calibration frames.

What is a flat field? A flat field is always best explained by looking at one. The image below shows a typical flat field. It is made by hanging some sheets over the end of the telescope and shining a diffuse light source at it. The CCD camera is then exposed for a short period of time (typically a second or so) so that values of around 20,000 ADU are obtained. . In the perfect world described above, your flat field would be just that - flat. A uniform white image with identical values on every single pixel. If your flat fields do look like this, then, unless the unlikely scenario above applies to you, you are probably doing something wrong. The most likely thing you are doing wrong is over exposing your images.

sample astrophotography flat field from ccd camera and telescope

Back to our real flat field complete with dark bits and knobbly bits. This represents what your CCD Camera and telescope "sees" when it looks at a flat, even light source. The main contributors are dust, scratches and vignetting.

Now, our favorite targets are not flat or even. They are faint, low contrast and hiding behind clouds. However, when you do capture some promising looking light frames, they are going to be afflicted by the same false dark bits and knobbly bits you see on your flat field. When you start stretching your image these artifacts cause problems. For example, you'll find that the corner which is dark on the flat frame will hit the black point before other corner, leaving an uneven mess. You wont' be able to increase the contrast on your target object as much as you would perhaps like.

These effects tend to ruin perfectly good astrophotography images.

So we take flat fields.Several in fact. How many is a subject of some debate, but, provided you have at least as many flat frames as you have light frames, then you'll probably be ok. We use our stacking program average the flat fields into a master flat frame, and then incorporate this information in the stack of light frames. Each light frame must be divided by the flat field to remove the irregularities, and then all he frames are stacked as normal - and nearly all of the artifacts vanish. Hooray.

Processing, especially stretching and contrast boosting becomes much much eaiser.

Anybody suffering problems in their images, and not taking flat frames because of some fake time constrain, or simply out of tirdness or apathy, is doing themselves and their astro images a huge injustice.

Some poeple get confused by the process of "calibrating" their images with bias frames, dark frames, flat frames etc. This is understaable, as their is a fair bit of terminology to grasp, however, it is not difficult to get an understanding. My friend Carsten Arnholm has a nice website explaining image calibration with the brilliant ImageViewCA software making the whole process simple.

What value should your flats be? Well, if you have a good quality cool CCD camera, then the flats should be approximately half the value of saturation. In most CCD cameras this is roughly 20,000 ADU.

There is another arguement for webcams and DSLRS. These devices are not scientific instruments, and as such, are not likely to be as linear as a dedicated astronomy CCD camera. The consquence of this is that the flats might not "work" perfectly. Robin Leadbeater has done a lot of work on webcam linearity and his pages explain the problem fairly well. Unless you are correcting the flats according to his system, it might be worth making DSLR and webcam flats with an ADU range similiar to the light frames. Just take more of them to reduce the noise.


There is no excuse for not flat fielding your astrophotograhy images.