Is there any solution to calculate with a QGIS expression the largest empty circle of a point layer as shown in the image below?
Maybe there is some solution combining or improving the functions pole of inaccessibility, convex hull in this way:
pole_of_inaccessibility(convex_hull(collect($geometry, 9999)), 9999)
I have tried this solution but the pole of inaccessibility concept is not the same as the largest empty circle concept.
Clarifications and possible solution:
A proposed solution would be the following:
Consider that the features extent would be defined by the detection of the most isolated points. This calculation can be done by reusing the following function that appears in one of the answers in the post Creating the minimum convex hull that contains certain percentage of points in QGIS:
array_sum ( array_foreach ( overlay_nearest( @layer, $geometry, limit:=2 ), length ( make_line ( $geometry, @element ) ) ) )
The next step would be to find the Minimum Delimiting Geometry with circle geometry formed by the most isolated point. This can be represented using Geometry Generator and the following function:
case when "DISTA" > mean("DISTA") then minimal_circle( collect_geometries ( overlay_nearest ( 'PG', $geometry, limit:=2 ) ) ) else $geometry end
Finally, we would have to manually eliminate the cases that do not correspond to the desired objective.
I think it is an unelegant solution and I am still not convinced by the fact that the last step is manual.
Could someone provide an improvement to this view or, perhaps, a new approach?




