how to expand the polygon to reach nearby line in r?

Question

Let say i have a SpatialPolygons object with 3 polygons data name groupexc:

library(raster) 
p1 <- matrix(c(2, 3, 4, 5, 6, 5, 4, 3, 2, 4, 5, 6, 5, 4, 3, 2, 3, 4), ncol=2)
p2 <- matrix(c(8, 9, 10, 11, 12, 11, 10, 9, 8, 4, 5, 6, 5, 4, 3, 2, 3, 4), ncol=2)
p3 <- matrix(c(5, 6, 7, 8, 9, 8, 7, 6, 5, 9, 10, 11, 10, 9, 8, 7, 8, 9), ncol=2)
groupexc <- spPolygons(p1, p2, p3)

And a SpatialPolygons object zoneexc that represents a single zone:

zoneexc = spPolygons(matrix(c(2,1,3,4,6,8,10,13,14,14,12,10,8,6,4,2,1,3,7,10,12,14,12,6,4,3,1,1,1,1,1,1), ncol=2))

Is there a way for me to expand the output from groupexc until it reach points in zoneexc?

before

plot(zoneexc, border='red', lwd=3)
plot(groupexc, add=TRUE, border='blue', lwd=2)
text(groupexc, letters[1:3])

after:

Any help would be appreciated.

please provide the data as code like `groupexc <- cbind(c(1,2,3), c(1,2,3))` (and see `raster::spPolygons` to make polygons.) — Robert Hijmans, Jun 10 '16 at 01:06
i didnt need any specific condition.myb they can expand fairly until they reach point to `zoneexc`. i just need it to be expand so that there is no hole in my polygon — ahmad fikri, Jun 10 '16 at 04:07
no sorry was wrong.i just want to expand it until its exterior part — ahmad fikri, Jun 10 '16 at 04:15

Robert Hijmans · Accepted Answer · 2016-06-16T18:12:05.663

3

Here is an approximate solution. This approach might break for large problems, and it depends on having sufficient number of nodes in each polygon. But it may be good enough for your purpose.

# example data 
library(raster) 
p1 <- matrix(c(2, 3, 4, 5, 6, 5, 4, 3, 2, 4, 5, 6, 5, 4, 3, 2, 3, 4), ncol=2)
p2 <- matrix(c(8, 9, 10, 11, 12, 11, 10, 9, 8, 4, 5, 6, 5, 4, 3, 2, 3, 4), ncol=2)
p3 <- matrix(c(5, 6, 7, 8, 9, 8, 7, 6, 5, 9, 10, 11, 10, 9, 8, 7, 8, 9), ncol=2)
groups <- spPolygons(p1, p2, p3, attr=data.frame(name=c('a', 'b', 'c')))
zone <- spPolygons(matrix(c(2,1,3,4,6,8,10,13,14,14,12,10,8,6,4,2,1,3,7,10,12,14,12,6,4,3,1,1,1,1,1,1), ncol=2))

Now create nearest neighbor polygons. For this to work as below, you need dismo version 1.1-1 (or higher)

library(dismo)
# get the coordinates of the polygons    
g <- unique(geom(groups))
v <- voronoi(g[, c('x', 'y')], ext=extent(zone))
# plot(v)
# assign group id to the new polygons
v$group <- g[v$id, 1]

# aggregate (dissolve) polygons by group id
a <- aggregate(v, 'group')
# remove areas outside of the zone
i <- crop(a, zone)

# add another identifier 
i$name <- groups$name[i$group]

plot(i, col=rainbow(3))
text(i, "name", cex=2)
plot(groups, add=TRUE, lwd=2, border='white', lty=2)

To see how it works:

points(g[, c('x', 'y')], pch=20, cex=2)
plot(v, add=TRUE)

edited Jun 16 '16 at 18:12

answered Jun 12 '16 at 03:18

Robert Hijmans

40,301
4
55
63

whats the -999 for? – ahmad fikri Jun 13 '16 at 00:51
this is work with the data. but when i am using my data, it says, Error in `[[<-.data.frame`(`*tmp*`, name, value = c(1, 1, 1, 1, 1, 1, : replacement has 8400 rows, data has 8395). do u have any idea why its become like this? – ahmad fikri Jun 13 '16 at 06:18
it happen at here, v$group <- g[,1]. i will try to learn to make my data available via fileshare. i actually didnt know how to share my data – ahmad fikri Jun 14 '16 at 00:31
and after i do geom, the variable name for a,b,c is gone. how can i keep it? – ahmad fikri Jun 14 '16 at 06:48
i think it is because like u said it will have problems if data is big. i can run it smoothly by using smaller sample. – ahmad fikri Jun 15 '16 at 01:04
where can i get your email? – ahmad fikri Jun 15 '16 at 03:28
i didnt get what u get. Error in voronoi(g[, c("x", "y")], ext = extent(zone)) : unused argument (ext = extent(zone)) – ahmad fikri Jun 17 '16 at 01:49
Then you are not using dismo version 1.1-1 – Robert Hijmans Jun 17 '16 at 01:55

how to expand the polygon to reach nearby line in r?

1 Answers1