Table of Contents


ccdfill - patch up the holes in an image


ccdfill [parameter=value]


ccdfill patches up the holes in an image.

Currently a simple 2-dimensional planar fit is done to the pixel values surrounding ’bad’ pixels. The size of this area can be choosen by the user.

This procedure will obviously not conserve flux (see also all=t below), and has to be used with caution.


The following parameters are recognized in any order if the keyword is also given:
Input image(3NEMO) file. No default.
Output image file. No default.
Number of neighbor cells at one side to use in interpolation/fit. The total number of cells used in the fit will hence be at most 4*n*(n+1) (this already excludes the central pixel, n=1 uses 8 cells, n=2 uses 24 etc.) [Default: 1].
Value in the map that should be patched. [Default: 0.0].
Normally only bad pixels are patches. By setting all=t all pixels are re-fitted with a planar fit, effectively smoothing the map. This will obviously not conserve flux, where applicaple, and should be used with caution. Default: f
Minimum number of neighbors needed to allow a fit. Normally 3 is enough, though this tends to add to edges. 4 would be more conservative, in which case iter>1 can be used. Default: 4
Minimum number of iterations. 0 is also an allowed value, in which case the image will just be copied unmodified. Default: 1
If given, for each point in the map, if a point deviates more than spike from all it’s 4 non-bad direct neighbors, this value will be reset to bad before the actually filling will take place. Default: none (not applied)


With m=3 you will see some singular matrix warnings, and patched values will be bogus. It is better to set m=4 for those cases.

Edge effects will occur for large values of iter, use with caution. Really a better bi-linear interpolation is needed. See also ccdintpol(1NEMO) if your irregulaties are regular enough.

See Also

image(3NEMO) , ccdmath(1NEMO) , ccdsmooth(1NEMO) , ccdintpol(1NEMO)


Peter Teuben

Update History

10-Dec-93    V1.0 Created    PJT
22-feb-97    V1.1 added bad=    PJT
19-feb-02    V1.2 added all=, m=    PJT
3-may-03    V1.3 added iter=    PJT
3-may-2017    V2.0 added spike=    PJT

Table of Contents