--nodata-radius with ipfind... unusually long run-times

[jchollingsworth]

I noticed that using ipfind with --nodata-radius option increases the run time by orders of magnitude.

The end result is what I would expect. But I wonder if there is a problem with how the filtering is implemented? If points are removed within the specified pixel radius after all IPs have been found, then I would expect that to be a near instantaneous operation(?).