Douglas-Peuker line generalization for Spatial Polygons

The function applies the implementation of the Douglas-Peuker algorithm for line generalization or simplification (originally from shapefiles) to objects inheriting from Spatial Polygons. It does not preserve topology, so is suitable for visualisation, but not for the subsequent analysis of the polygon boundaries, as artefacts may be created, and boundaries of neighbouring entities may be generalized differently. If the rgeos package is available, thinnedSpatialPolyGEOS will be used with partial topology preservation instead of the R implementation here by passing arguments through.

thinnedSpatialPoly(SP, tolerance, minarea=0, topologyPreserve = FALSE,
                 avoidGEOS = FALSE)

Arguments

SP	an object inheriting from the SpatialPolygons class
tolerance	the tolerance value in the metric of the input object
minarea	the smallest area of Polygon objects to be retained, ignored if rgeos used
topologyPreserve	choose between two rgeos options: logical determining if the algorithm should attempt to preserve the topology (nodes not complete edges) of the original geometry
avoidGEOS	use R DP code even if rgeos is available

Value

An object of the same class as the input object

References

Douglas, D. and Peucker, T. (1973). Algorithms for the reduction of the number of points required to represent a digitized line or its caricature. The Canadian Cartographer 10(2). 112-122.

Author

Ben Stabler, Michael Friendly, Roger Bivand

Note

Warnings reporting: Non-finite label point detected and replaced, reflect the changes in the geometries of the polygons induced by line generalization.

Examples

xx <- readShapeSpatial(system.file("shapes/sids.shp", package="maptools")[1],
      IDvar="FIPSNO", proj4string=CRS("+proj=longlat +ellps=clrk66"))
#> Warning: readShapeSpatial is deprecated; use rgdal::readOGR or sf::st_read
#> Warning: readShapePoly is deprecated; use rgdal::readOGR or sf::st_read
object.size(as(xx, "SpatialPolygons"))
#> 350544 bytes
xxx <- thinnedSpatialPoly(xx, tolerance=0.05, minarea=0.001)
object.size(as(xxx, "SpatialPolygons"))
#> 374480 bytes
par(mfrow=c(2,1))
plot(xx)
plot(xxx)
par(mfrow=c(1,1))