aswheaton/fits-utils

Centroiding function fails to get fix on reference star

Closed this issue · 2 comments

Post all relevant troubleshooting and trial fixes here. Relevant functions are:

weighted_mean_2D()
max_value_centroid()
threshold_centroid()

Compared the way that pyplot.imshow and SAO DS9 index arrays with some test code to make sure our cutout wasn't excluding the reference star due to different indexing conventions. Here is an array of zeros with shape (100,100) and indices [0:50,0:50] set to one in pyplot.imshow:
imshow
And the same array exported to DS9:
ds9
As you can see, imshow puts, by default, the origin in the upper left corner, but both tools place the modified quarter in the corner associated with the origin, so indexing conventions should not be the issue here.

Fixed by the addition of hybrid_centroid() function in commit a692062.