Home

• Photoshop
  • Mosaics
• Iris
  • Preprocessing

Preprocessing using Iris

This tutorial leads you through the basic steps of applying darks, flats and bias frames to your image, aligning your images and stacking them in Iris. It can be quite a long and arduous process, so I hope this tutorial can make it quite a lot easier!

• Step 1
  • Setting up Iris needs you to tell it where you want all your images saved
  • Go to File > Settings... and the window shown below will appear
  • Change the 'Working Path' to where you want all your images to be saved and then click 'OK'
  • You also need to tell it which camera you have so you are using the right file type. Click the camera icon (3rd in from right) and you will see a drop-down box

• Step 2
  • You now need to select the images you are going to use
  • Click Digital Photo > Decode RAW files...
  • Iris will now dissappear and be replaced by a smaller window as shown below
  • Using whatever method you prefer, find the images on your computer (e.g. My Documents > My Pictures etc.)
  • Now select only the light frames. Click and drag them into the Iris window. Give them a sensible generic name such as 'M45', and then click 'CFA...'
  • Repeat the previous step for all your flat, dark and (if you have them) bias frames
  • Click 'Done' when you are finished and the main Iris window will reappear

• Step 3
  • If you didn't take any bias frames then you need to follow this step. Otherwise, skip this step!
  • We are now going to create a 'fake' bias frame
  • Load one of your converted images. Click File > Load. Because you gave your images generic names they will be called M451, M452 M453 etc. Select any one and click 'Open'.
  • Now open the 'Commands Window'. It is the 4th button from the right on the toolbar, just to the right of the camera icon (you can just about see it in the image below).
  • A small window will appear. Click on the first line and type 'fill 0' (without the apostrophes). Hit enter and the whole image will go black. Now type 'save bias'. This is now your bias frame.

• This is what your 'command' window should look like after saving

• Step 4
  • If you did take bias frames then you need to follow this step. Otherwise skip it and go onto step 5!
  • To create your bias frames, click Preprocessing > Make an Offset...
  • The window is pretty straightfoward. Next to 'generic name' type the generic name you called your bias frames, and then next to 'number' specify how many you saved. Click 'OK'.
  • Now save the bias frame in the same way as the bias was saved in Step 3.

 

• Step 5
  • You will now create your dark frame
  • Go to Preprocessing > Make a Dark...
  • The following dialogue box will appear - fill it in. The generic name is what you saved your darks as in Step 2, the offset image is your bias frame which you created in Step 3 or 4 and the number is how many you saved.
  • I find that the 'Median' stack works best, but you can experiment!
  • In the same way as before (using the 'command' box, save the dark as 'dark' (or something similar))

• Step 6
  • You will now create your flat-field
  • Follow the same steps above to fill out the dialogue box
  • The only difference is the 'Normilization value' field. There is no magic number to go in there. After you have clicked 'OK' you will see the resultant flat-field. If it looks too bright or dark to you, redo this step and adjust the Normilization value until the result looks ok. You will probably need two or three goes to get it right!
  • In the same way as before (using the 'command' box, save the flat as 'flat' (or something similar))

• Step 7
  • This step finds the hot pixels in your image to improve dark-frame subtraction
  • Firstly, open your master dark by clicking File > Load and then 'Open'
  • Click the command box. Type 'find_hot cosme 500' (without the apostrophes) and then press enter
  • An 'output' box will appear as shown below. It will tell you the number of hot pixels found. You want something between 100 and 200
  • If you are below 100, change the 500 in find_hot cosme to something smaller
  • If you are above 200, change the number to something larger

• Step 8
  • The preprocessing step.
  • To open the Preprocessing window, click Digital Photo > Preproessing...
  • The following box will appear - fill it in!
  • The input generic name wants the generic name of your lights, and the number box wants to know how many you have got
  • The output generic name box is what it will save the resultant images as. I usually put a _pp after the generic name of my lights, so in this case I have 'M45_pp'
  • Don't click the optimize tab
  • Click 'OK'. The process shouldn't take too long, but it depends on the number of images you have. A good guide is around 1 minute per image (computer dependant), so if you have 30 then do something else for a while!

• Step 9
  • You are now going to convert your images to colour
  • Go to Digital Photo > Sequence CFA Conversion...
  • The box below will appear - fill it in!
  • Enter the name of your preprocessed lights, and enter a suitable output name (I stick _clr on the end so in this case it is 'M45_clr')
  • Tell it how many lights you are using, make sure the 'colour' box is checked and click 'OK'. The process doesn't take too long (it is a lot faster than the preprocesing)

• Step 10
  • Aligning your images. This step can be really processor intensive, so I would close any other programs you are using
  • Click Processing > Stellar registration and the box below will pop up
  • As before, fill it out - tell it the input lights, give it a sensible output name (I use M45_rgd) and tell it the number of lights you are using
  • Now you need to decide what aligning process you want to use. Personally I prefer the 'Global matching' process, without selecting a zone. The transformations do all sorts of cunning things to make your images align. Quadratic seems to work well, but you will sometimes want to try the others. Basically quadratic will rotate, move, shift and sometimes warp your images to match.
  • Another method, which is faster is the 'Three Matching Zone' method. All you need to do is enter a suitable Zone size and click 'OK'.
  • Click 'OK'. You might want to check it is ok every few minutes because sometimes it doesn't like an image. If this happens, either remove it from the images you want to align, or choose a different aligning method.

• Step 11
  • You are nearly finished! This step stacks all your images
  • Click Processing > Add a sequence and the window below will appear
  • As before, fill it in! There is no best method to use, but I have found that:
  • Sigma clip, Sigma median and arithmetic are the best three. In this case I used Sigma clip
  • The 'Sigma coefficient' and 'Number of Iterations' boxes have preset values, but you might want to change them. As before, run it first with the preset values, and if you don't like the result, change them.
  • Finally, click 'OK'. The process shouldn't take too long. To save your image, click File > Save and save it where you like in the format you want.

• The result!
  • After a bit of touching up in Photoshop (using guess who's actions :-) ), this is the final result!