Processing steps for astrophotographs


This page is up to date for GRIP version 12.3.15 but see the "What's new" page for more recent changes.

This process is available on the Batch/Astro menu of GRIP. There are now (2012 March) several relevant options on that menu. Read about them here and then come back to this page.

Two of the options ("Astro combine into 1 image" and "Astro warp/shift onto common basis") cause the astro-process dialogue to be displayed. Its title shows which option was selected: the title is either "Astro COMBINE" or "Astro ALIGN".

 Controls in the astro-process dialogue


 Dark frames

The original plan was to identify dark frames automatically among the sequence of images and then, for each real image, to subtract the dark frame which was taken at the nearest time. The identification of dark frames proved to be unreliable so a simpler approach has been adopted. It is now assumed that you have used the averaging batch process on a series of dark frames beforehand. The dialogue explained above asks whether there is such an averaged dark frame and where it is.

There is more about dark frames here.


 Flat fields

It is possible to use flat field images, if available, for the process. A flat field is a photo of a uniformly illuminated white area, completely filling the frame. This must be photographed with exactly the same optical arrangement as the images for which this is to be used to correct.

There is more about flat fields here.

 DSLR darks and flats

With a DSLR it is not really necessary to use dark and flat frames during the stacking process. For one thing, hot pixels can be mapped out in the camera. (For Canon cameras, before each session do the following. Put the body cap on. Select the menu option for manual sensor cleaning, which flips the mirror up. Leave it like that, mirror up and body sealed, for about a minute. Then switch off, releasing the mirror. At this point it seems that the camera stores a list of hot pixels. This list must go into the metadata of each image so that Adobe Camera Raw and similar programs can use it when interpolating for the Bayer filter. I have not been able to work out where the list is, to use it in GRIP. Instead I convert my RAW images to TIFF before stacking, using Canon's DPP application that came free with the camera, and that process apparently uses the list.)

The flat field can be applied to the result after the stacking as a manual step on the image combination menu, rather than applying it to every frame along the way - unless there is very large movement of stars through the image sequence.

 The steps of the process

  1. User selects an arbitrary number of image files to process. They must be all in the same format but that may be RAW if the particular RAW format from the camera can be read by the jrawio library (find that out by trying to open a RAW image file in GRIP before starting this process; if it fails it will be necessary to convert the images to TIFF through other software before starting this process)
  2. The user selects an averaged dark frame image (if there is one); it is loaded; if it is RAW it is interpolated, scaled and its dark margin trimmed (as set in the configuration menu)
  3. The user selects a flat field image (if there is one); it is loaded; if it is RAW it is interpolated, scaled and its dark margin trimmed (as set in the configuration menu)
  4. Pass 1: for each image file
    1. load the image from file (into memory but do not waste time displaying it)
    2. if the image was in RAW format, do the necessary interpolation and scaling and trim its dark margin (as set in the configuration menu)
    3. if there is a dark frame, subtract it
    4. if there is a flat field image, divide by the flat field image.
    5. if selected by the user, correct background
    6. if gnomonic projection selected by the user, find lens focal length mm from metadata and do inverse gnomonic projection
    7. if the image has been modified by any of the preceding steps, save the adjusted image as a temporary file in TIFF format
    8. segment stars in the image, as blobs (contiguous groups of pixels)
    9. list the positions of the centres of the brightest N objects (N is a configured parameter, default value 24; see the configuration menu)
  5. Analyse the positions measured in all the images to match brightest blobs between images
  6. Pause to let the user reject some matched objects (not usually necessary, but in case some non-star objects have been matched). This is done by displaying a diagram of all the matches and asking the user to draw rectangles intersecting the match trails which are to be rejected. Here is a portion of an example of such a diagram (the diagram is displayed as an image in GRIP's usual way, so it can be manipulated and saved also).
    This diagram shows part of the constellation of Andromeda. Numbering starts at 0 and roughly increases as measured star brightness decreases (though it is really more complex than that). Confirmation is sought before each trail is deleted.
    It is necessary to close the window containing the diagram in order to proceed to the next step.
  7. Another window opens at the end of each pass of the batch astro-process to summarise the displacements between all the images that have been processed. An example can be seen here.

  8. Pass 2: for each image
    1. load the image from file (from the temporary file if one was made in pass 1)
    2. use the measurements from pass 1 to warp the image slightly, like a rubber sheet, so its brightest objects fit over the brightest objects in the middle image in the sequence
    3. add the warped image into an accumulator image (32 bits per channel)
    4. delete the temporary file, if there is one
  9. Save the result as a FITS file, 32 bits (integer) per channel, as a safeguard in case the user forgets to save it.
  10. Display the result in the usual GRIP image frame with all menus available.
Next page