Image Calibration Reference

 

This is a quick and dirty reference to calibrating your astronomical images.

 

Why calibrate? Calibration helps remove unwanted artifacs from your images, potentially making them prettier. However, the primary reason for calibration is accuracy and consistency in the image. This is -extremely- important if you are doing scientific work. Such as photometry measurements, exoplanet searches, and the like.

 

This is not much of a tutorial, but rather a set of notes to help guide when processing images.  There are tutorials online, as well as publications such as AIP4WIN through William Bell publishing which can fill you in on all the details.

 

I do not claim to be 100% accurate here, nor do I claim that these procedures will work for everyone at all times.  These procedures were for the most part compiled by myself through various sources. The mentioned tutorials, books, and from organizations such as the AAVSO. I will fine tune this document over time as I tweak/tune my methods.

 

ACCURACY (the 50% well depth rule)

For photometry, when using an ABG (anti-blooming) camera, it is suggested not to exceed 50% pixel ADU (well depth), as CCD's can lose linearity beyond this point. Know your camera well if are to do scientific work. On some newer cameras linearity isn't quite as badly affected by ABG. For good measure however, the 50% rule is a safe bet. If you are not doing photometry or some other scientific measurement, then this limit is not quite as important.

 

accuracy= ~<50% pixel ADU for ABG cameras.

     <>SX MX716 max ADU = >70,000e-              (1x1 binning)

     <>Effective linear well depth = ~35,000e-      (1x1 binning)

 

CALIBRATION EQUATIONS

 

MasterFLAT = (RAW - DARK) / FLAT

CalibFrame = (RAW - BIAS - DARK) / MasterFLAT

 

BIAS (readout noise)

Bias image is 0 Second Exposure. Take as many as possible.

 

definition: Measurement of readout noise of CCD system. Noise can vary based on environmental factors. Nearby equipment emitting RFI, power quality, equipment used, etc. The more you take, the better. Different CCD's introduce a degree of readout noise. Typically the older, the noisier.

 

IMAGE PATTERNS = NOISE (i.e. vertical blinds)

SX MX716 USB1 readon noise is quite small at <8e-

noise decreases by SQRT(X) where X=#Frames

General Rule: Take X Bias images for X Darks.

 

COMBINING BIAS FRAMES  -> MASTER BIAS

    <>If in a noisy environment -> use Median Combine 

    <>Otherwise ->  use Average Combine

 

DARKS (thermal readout)

IMAGE-TIMES-FIVE-RULE: suggested typical ratio. For a 2 min image/frame, take 5x 2minute darks.  OR 10x 1 min darks (to be scaled). Cover the CCD detector, expose for time X.

 

DARK ALONE = NOT SCALABLE

DARK - BIAS = IS SCALABLE

 

definition: A dark is a measure of the CCD's thermal state. This dark baseline can vary depending on environmental temperature. The warmer the environment, the noisier they will be. Darks may be archived and  reused for future projects when the environment is the same temperature. While convenient, it is best to take fresh darks with each imaging session. Check each dark for cosmic ray hits before use (white streaks).. it is best to discard them.

 

Some cameras have such a low dark current, that darks can potentially be omitted. While this is Ok for more casual or pretty pictures, it is still highly suggested when doing scientific work.

 

COMBINING DARK FRAMES -> MASTER DARK

     <>Subtract Master BIAS from EACH individual Dark FIRST.

     <>For highest quality ->  Average Combine

     <>For 3 dark frames minimum -> Median Combine (more noise!)

 

FLATS (subtracting optical imperfections in light path)

Perfecting flats is more of an art than a science. Obtaining good/great flats can be a real challenge. I won't go into flats much, or how to specifically create them.  There are documents out there on the topic. What is good for someone else, might not be good for you however. For proper flats, you need both Flats and Flat Darks. You can omit the darks, but if you are lazy about it, you might as well skip flats and lose that extra level of accuracy in your measurements.

 

FLAT DARK EXPOSURE = FLAT EXPOSURE

     -Expose until 50% well depth. IMPORTANT

     -16x exposures requirement, MINIMUM

     -Keep SNR @ 500:1 for 0.1 mag photometric accuracy.

 

     <>Average FLATS

     <>Average or Median Combine DARKS for FLATS (see darks section)

 

MasterFLAT = AverageFLAT - AverageDARK

     *divide flat into dark subtracted image*

MasterFLAT = (RAW - DARK) / FLAT

 

HANDY LINKS:

 

Starlight Xpress Camera Specs

http://www.starlightccd.com/prodserv/overview.htm

MX-716/7C specific information

http://www.astroturtle.com/mx7.htm