• 1.1 Manual Calibration Procedure

• 1.2 Automatic Calibration

Image calibration uses support frames to remove unwanted fixed signals (such as thermal current and bias), and corrects signal modifications (such as vignetting) so that the raw image in the light frames accurately represents the intensity of light incident on the sensor during the exposure.

Calibration may or may not be necessary for every image. This will depend on several different factors:

• The level of excellence desired in the final image.
• The amount of unwanted signals and modified signals present in the light frames.
• The amount of work you are prepared to perform.

Basic and Advanced Calibration

There are two different kinds of calibration that you can apply to the light frames.

• Basic Calibration - Dark frames are used to remove the thermal signal from the light frames. Since bias is already present in every frame, including the darks, subtracting a dark frame also automatically subtracts a bias frame. Flat-fielding is not performed.

• Advanced Calibration - A master bias frame is subtracted from a master dark frame to create a scalable master thermal frame. The master thermal frame is used to remove the thermal signal from the light frames. The master bias frame is used to remove the bias signal from the light frames. Flat-field frames are used to correct vignetting, dust shadows, uneven illumination and unequal pixel response.

Many astronomical image processing programs will automatically perform these calibration steps once you specify which files are to be used.

Advanced image calibration needs to be done in an astronomical image processing program like Images Plus, Iris, or Maxim DL. Photoshop cannot perform the division necessary for flat-fielding, or scaling of the master thermal frame that most of these programs do automatically.

Frames Needed for Basic Calibration

To do basic calibration with you need the following types of frames:

• Light Frames - exposures of the deep-sky object.

• Dark Frames - shot at the same exposure time, ISO and temperature as the light frames but with no light reaching the sensor.

Frames Needed for Advanced Calibration

To do complete advanced calibration with a scalable dark frame you should have the following types of frames:

• Light Frames - exposures of the deep-sky object.

• Dark Frames - shot at the same exposure time, ISO and temperature as the light frames but with no light reaching the sensor. To be scalable, the exposures for the individual dark frames should be longer than the longest exposure you plan to use for your light frames. A dark frame is a picture of the camera's thermal current.

• Bias Frames - shot at the same ISO and temperature as the light frames at the shortest exposure of which the camera is capable with no light reaching the sensor. A bias frame records the bias signal present in every frame. Bias is a low-level charge that is applied to a CCD or CMOS sensor in the form of a fixed offset voltage value.

• Flat-Field Frames - an evenly illuminated featureless frame shot at a low ISO on automatic exposure. A flat-field frame is used to correct for vignetting, uneven illumination, dust on the sensor's cover glass, and uneven pixel response.

Calibration Before De-Bayerization

The Bayer pattern of individual red, green and blue colored filters over each pixel technically produces a "Bayerized" black and white image. When this image is "de-Bayerized", color is produced. Because de-Bayerization involves interpolation, the original information is changed when data is interpolated. This information includes all of the signals, wanted and unwanted, and noise in the image. Because of these changes, calibration is not optimum if it is performed on the light frames after they have been de-Bayerized.

For example, there may be individual hot pixels in the thermal signal. If the image is calibrated before de-Bayerization, this individual hot pixel will be correctly removed. If the image is calibrated after de-Bayerization, this individual hot pixel may be interpolated into something different in the light frame than what it will be interpolated into in the dark frame, because color information is present in the light frame that is not present in the dark frame.

Red, green and blue color "blobs" are also present in images with a high thermal signal. These blobs are better reduced or eliminated if calibration is done before de-Bayerization and color creation.

Image calibration should be performed on the light frames before they are de-Bayerized and color created.

Linear vs Non-Linear Files

Just as with de-Bayerization, the original data in a raw file is linear. Calibration should be performed on the original raw linear files before de-Bayerization.

Almost all raw converters that are used with normal photographic images, such as the programs that come with the camera from the manufacturer like Canon's Digital Photo Professional, and Nikon View, and even most third-party image processing programs like Photoshop, will apply a default non-linear stretch when a raw file is initially opened. It is difficult to get a true linear file out of these programs. It is much better to use an astronomical image processing program like Images Plus, AIP, IRIS, or Maxim to process the linear 16-bit raw files from DSLR cameras.

The bottom line is that calibration should be performed on the raw linear files in 16-bit depth before the image is de-Bayerized.

Frame Components

Each type of frame is made up of different components, including noise. Since noise is random and non-repeatable, noise will not be included in the following definitions, but it is present in every frame. The only way to deal with noise is to increase the signal.

• Light Frame = (Object Signal * Flat Signal) + Thermal Signal + Bias Signal

• Dark Frame = (Thermal Signal + Bias Signal)

• Flat-Field Frame = (Flat-field Signal + Flat-field Thermal Signal + Bias Signal)

• Bias Frame = Bias Signal

Technically, because the bias frame in a DSLR camera is not a zero-second exposure, it has a thermal signal also. But because it is an extremely short exposure, it will be negligible and we don't have to worry about it.

Likewise, flat-field frames are generally shot at a sufficient level of illumination so that the exposures are short enough so that the thermal signal in the flat is negligible, but it is present. For all practical purposes, you don't have to shoot a thermal frame for your flat-field frame, unless the exposure for the flat gets very long, or you want to be especially accurate.

Because the temperature increases in a DSLR over the course of multiple sub-exposures, the thermal signal increases in every frame. This makes it difficult to exactly match a thermal frame to a light frame. It is probably the best idea to use a scalable master dark frame with Images Plus Adaptive Dark Subtraction. Images Plus will analyze each frame and calculate the optimum amount of scaling for the dark frame.

If you are going to use a scalable master dark frame, it is important to shoot bias frames. Since the scalable master dark frame is stretched, the bias present would be stretched also, and it would not match the bias in the light frames anymore. If scalable dark frames are used, bias frames should be subtracted separately from both the light frames and the dark frames.

A Calibrated Image = (Light - Bias - Dark) / Flat