如何为 R 中的点坐标构建 Voronoi 多边形？

Question

I have various points (2000+) for observation stations in the alps. I would like to use them to represent the closest geographic area, that is not closer to another observation station. I have done some research, and think that using Varanoi polygons may be the best way to do this.

After having attempting to build these in R, the polygon plot does not quite match my graphing in R.

I have attached the sample data points I am experimenting with, as well as the two images that show the dissimilar graphing of the points.

What do I need to be do differently to make sure that the points line up?

Points:

Longitude:
15.976667 12.846389 14.457222 13.795556  9.849167 16.055278 13.950833 15.666111  9.654722 15.596389 13.226667 15.106667 13.760000 12.226111  9.612222 17.025278  9.877500 15.368056 13.423056 12.571111 16.842222 13.711667 14.003056 12.308056 13.536389

Latitude:
48.40167 48.14889 47.56778 46.72750 47.45833 48.04472 47.82389 47.49472 47.35917 48.64917 48.25000 48.87139 47.87444 47.42806 47.20833 47.77556 47.40389 47.87583 47.53750 46.77694 47.74250 46.55000 48.37611 47.38333 47.91833

Pictures: Map of the 25 sample points in Leaflet:

Voronoi plot:

Clearly these two are not the same images, so I must be doing something wrong. Here's the code I'm using to generate the Voronoi plot and the leaflet map.

meta25%>% 
  st_as_sf(coords = c("Longitude", "Latitude"),
           crs = sp::CRS("+proj=longlat +datum=WGS84")) %>% 
  mapview()

m1 = matrix(meta25$Longitude,meta25$Latitude,ncol=2,nrow=25) %>% st_multipoint()
voronoi_grid <- st_voronoi(m1)
plot(voronoi_grid, col = NA)
plot(m1, add = TRUE, col = "blue", pch = 16)

Answer 1

I'm not sure what the problem is, but the matrix is not necessary. Stick to sf objects and you should be fine.

library(tidyverse)
library(sf)


# create pts from lat & lon data
pts <- tibble(latitude = y, longitude = x) %>%
  st_as_sf(coords = c('latitude', 'longitude')) %>%
  st_set_crs(4326)

# voronoi of pts
vor <- st_voronoi(st_combine(pts))

head(vor)
#> Geometry set for 1 feature 
#> Geometry type: GEOMETRYCOLLECTION
#> Dimension:     XY
#> Bounding box:  xmin: 2.199166 ymin: 39.13694 xmax: 24.43833 ymax: 56.28445
#> Geodetic CRS:  WGS 84
#> GEOMETRYCOLLECTION (POLYGON ((2.199166 49.37841...

# st_voronoi returns a GEOMETRYCOLLECTION,
#  some plotting methods can't use a GEOMETRYCOLLECTION.
#  this returns polygons instead
vor_poly <- st_collection_extract(vor)

head(vor_poly)
#> Geometry set for 6 features 
#> Geometry type: POLYGON
#> Dimension:     XY
#> Bounding box:  xmin: 2.199166 ymin: 39.13694 xmax: 18.32787 ymax: 56.28445
#> Geodetic CRS:  WGS 84
#> First 5 geometries:
#> POLYGON ((2.199166 49.37841, 2.199166 56.28445,...
#> POLYGON ((9.946349 39.13694, 2.199166 39.13694,...
#> POLYGON ((18.32787 39.13694, 11.64381 39.13694,...
#> POLYGON ((9.794868 47.23828, 9.766296 47.38061,...
#> POLYGON ((5.225657 56.28445, 9.393793 56.28445,...

plot(pts, col = 'blue', pch = 16)
plot(vor_poly, add = T, fill = NA)

^{Created on 2021-04-05 by the reprex package (v0.3.0)}

Answer 2

Thanks everyone for your help, not sure if it got quite to where I was looking for. I've since adapted the answer from here: Creating bordering polygons from spatial point data for plotting in leaflet