Run delta-ball analysis to obtain an outline points for
delta-ball covering where points are in 2d space. We first find the minimum
delta such that all balls centered at each point in the data set is
touching at least 1 other ball (see get_delta for more
information). This function then creates a geometric objects that trying to
represent the covering of all these delta balls. Specifically we use
geometric properties to find the points on the "outside" of this covering
and return a set of lines that create a "boundary" of shorts. Intuition
from "Computing Polygonal Surfaces from Unions of Balls" by Tam and
Heidrich was used in this function.
inner_delta_ball_wrapper( data_raw, xy_columns = c("x", "y"), n_steps = 100, remove_duplicates = F )
| data_raw | data frame with center points of balls |
|---|---|
| xy_columns | columns of data.frame that relate to
the points's coordinates in euclidean space. The input should look like
something like |
| n_steps | number of equidistance points along the line, past delta on both sides, that will be checked to approximate all points along the line |
| remove_duplicates | boolean if need to remove duplicates in data_raw |
data frame of exterior lines (not ordered)
This function (renamed as delta_ball_wrapper) is shared with TCpredictionbands on github:
TCpredictionbands.