Slice a CTDF into putative clusters using temporal continuity and spatial clustering

Identifies spatially heterogeneous regions (via HDBSCAN on point coordinates) and, for those regions, recursively subdivides the track into temporally continuous movement segments. Subdivision continues until a region is spatially homogeneous (no evidence for multiple clusters) .

Usage

slice_ctdf(ctdf, nmin = 5, deltaT)

Arguments

ctdf

A CTDF object.

nmin

Integer; smallest size of a putative cluster.

deltaT

Numeric; maximum allowable time gap (in days) between segment endpoints for intersections to consider them continuous.

Value

Invisibly returns ctdf, with .putative_cluster updated in-place.

Details

This function updates a ctdf in-place.

Internally, candidate regions are queued. Regions that show evidence for multiple clusters are split by movement segmentation; otherwise they are retained as a single putative cluster.

See also

hdbscan

Examples

data(mini_ruff)
ctdf = as_ctdf(mini_ruff, s_srs = 4326, t_srs = "+proj=eqearth")
ctdf = slice_ctdf(ctdf)

data(pesa56511)
ctdf = as_ctdf(pesa56511, time = "locationDate", s_srs = 4326, t_srs = "+proj=eqearth")
ctdf = slice_ctdf(ctdf)
#>  ⠙ 23 segments processed [2s]
#>  ⠹ 32 segments processed [2.2s]
#>  ⠸ 36 segments processed [2.4s]
#>  ⠼ 44 segments processed [2.6s]
#>  ⠼ 46 segments processed [2.7s]