[task #16547] Avoiding NoiseChisel crash and improvements

[Mohammad Akhlaghi]

Follow-up Comment #1, task #16547 (group gnuastro):

Most of the problematic options have been fixed and I am still working on
completing them. For example if no tiles with good '--meanmedqdiff' are found,
NoiseChisel will increment it until at least one tile is found; or when there
are not enough tiles for outlier rejection, NoiseChisel will not do it, or
when the number of tiles for interpolation are not enough, it will use what
ever it has. All these changes from the values on the command-line are
ultimately written as header keywords (for the caller to be able to check if
anything has changed).

In the meantime, I have also found some good improvements (for example, there
were problems in outlier rejection and interpolation). Here is the output of
the current under-development branch on the SDSS M51 image of the second
tutorial.

(file #56131)


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/task/?16547>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/