BIGMEDIAN: Compute Median of a Large Number of FITS Images

BIGMEDIAN loads the 'dest' buffer with the median of images specified in the file given by the LIST=file keyword. The list file must contain the names of at least 3 images, in FITS format, with one image name specified per line of the input file. Optionally, the name of the file can be followed (on the same line) with a normalization constant and a zero point to be applied to each pixel in this input FITS file before the median is performed. For example, if the list file looks like this:

then the median of those frames will be taken from the raw input FITS values. If the input file looks like this:

then the median will be taken of image001, (image002/20.), and (image003-5)/20. Note that the specified zero point is SUBTRACTED from the image before computing the median; this makes it ideal to use for subtracting e.g. an overscan value.

In addition, you can do a pixel-by-pixel bias subtraction for each frame in the input list if you specify a FITS bias frame to subtract with the BIAS=file keyword. In this case, the bias frame will be subtracted before the median is performed.

The normalization of each frame can also be specified if there is a FITS card called MEAN in the FITS headers. If this card exists, the value of the card will be used to normalize the individual frames. Finally, if the NOMEAN keyword is specified, no normalization will be applied, regardless of the presence of a MEAN card or a value in the input list file.