MOPEX Online Manual

Overlap Modules: Med Filter

PURPOSE

This module performs a median background subtraction of the individual input images in order to set the frames up for Bright Object Detection. The output background-subtracted BCDs are taken as input by Detect (Outlier), which carries out the detection of bright pixels in the frame.

PARAMETER BLOCK

&MEDFILTER
 Window_X = 45,
 Window_Y = 45,
 N_Outliers_Per_Window = 500,
&END

INPUTS

Window X, Y: (int) the X, Y size in pixels of the window used to compute the median of the background. The default value is 45.

N_Outliers_Per_Window: (int) the number of outlier pixels (N) being rejected from the X*Y window when computing the median background. A rule of thumb to remeber when setting these parameters is that if you have a very crowded field, the fraction of rejected pixel would be higher than a less crowded field. A fraction of 1-2% is an acceptable number for most datasets. If N is set too high, the background will be under-estimated.

OUTPUTS

Generated Fits Files (*_minback.fits): The background subtracted, individual images.

DISCUSSION

This module is required if you plan to use Bright Object Masking, and carries out a median background subtraction before running the Detect (Outlier) module to detect the bright objects in the frame. The background subtracted images are not used for later mosaicking.

The program computes an asymmetrically skewed median for each pixel in the input image using a rectangular window of Window_X by Window_Y pixels. It is achieved by omitting the N_Outliers_Per_Window highest pixels from each median window.