11

I am trying to remove the intersect from Poly2 onto Poly1. Below are the coords of the 2 polygons.

> coords1
          [,1]     [,2]
 [1,] 992.0161 7462.531
 [2,] 950.1962 7413.532
 [3,] 902.3632 7421.207
 [4,] 706.1985 7472.378
 [5,] 654.0139 7711.760
 [6,] 657.5960 7726.950
 [7,] 786.2667 7844.380
 [8,] 789.5935 7842.925
 [9,] 981.7046 7498.659
[10,] 983.4246 7493.271
[11,] 990.6680 7469.888
[12,] 992.0161 7462.531
> coords2
         [,1]     [,2]
[1,] 930.3464 7607.313
[2,] 979.3528 7502.470
[3,] 865.8662 7484.337
[4,] 850.5665 7594.947
[5,] 930.3464 7607.313

Poly1<-Polygon(coords1)
Poly2<-Polygon(coords2)

The 2 polygons looks like thisenter image description here What I would like to get at the end is the dark blue polygon (Poly1) without the light blue (Poly2) area i.e., new eastern coordinates.

PolyGeo
  • 65,136
  • 29
  • 109
  • 338
GodinA
  • 489
  • 2
  • 6
  • 15

3 Answers3

8

Try the gDifference function from the rgeos package:

coords1 <- as.matrix(read.table(textConnection("
992.0161 7462.531
950.1962 7413.532
902.3632 7421.207
706.1985 7472.378
654.0139 7711.760
657.5960 7726.950
786.2667 7844.380
789.5935 7842.925
981.7046 7498.659
983.4246 7493.271
990.6680 7469.888
992.0161 7462.531
")))

coords2 <- as.matrix(read.table(textConnection("
930.3464 7607.313
979.3528 7502.470
865.8662 7484.337
850.5665 7594.947
930.3464 7607.313
")))

library("sp")
poly1 <- Polygon(coords1)
poly2 <- Polygon(coords2)
# create SpatialPolygons objects
p1 <- SpatialPolygons(list(Polygons(list(poly1), "p1")))
p2 <- SpatialPolygons(list(Polygons(list(poly2), "p2")))

library("rgeos")
res <- gDifference(p1, p2)
plot(res, col="blue")

gDifference result

rcs
  • 3,894
  • 1
  • 27
  • 30
5

The solution using rgeos::gDifference is much faster, but it returns a simple SpatialPolygons object.

If you are working with two SpatialPolygonsDataFrames and you want to keep the dataframe information, you can simply do this, and the result will be a SpatialPolygonsDataFrame:

library(raster)

res <- p1 - p2
rafa.pereira
  • 1,267
  • 17
  • 27
  • Using this solution, I get the error: "non-numeric argument to binary operator" – Ashe Mar 17 '17 at 15:13
  • @Ashe , type class(my_obj) to check whether both of your objects are SpatialPolygonsDataFrame – rafa.pereira Mar 17 '17 at 15:22
  • Indeed, I get SpatialPolygonsDataFrame as the class. – Ashe Mar 17 '17 at 15:25
  • Hmmm strange. I'm not sure why this would happen. Perhaps you should open a new question of how to Remove intersect of two SpatialPolygonsDataFrames while keeping the data info – rafa.pereira Mar 17 '17 at 15:48
  • 1
    I have fixed the issue. I needed to install the raster library for this solution to work. That addition to the answer would be helpful. Also, credit to this answer: http://gis.stackexchange.com/a/169597/93643 – Ashe Mar 17 '17 at 15:48
  • thanks for the heads up. I've included library(raster) to the answer. – rafa.pereira Mar 17 '17 at 16:12
0

If you want to use the sf package, you can use the st_difference() function.

library(sf)

coords1 <- as.matrix(read.table(textConnection("
992.0161 7462.531
950.1962 7413.532
902.3632 7421.207
706.1985 7472.378
654.0139 7711.760
657.5960 7726.950
786.2667 7844.380
789.5935 7842.925
981.7046 7498.659
983.4246 7493.271
990.6680 7469.888
992.0161 7462.531
")))


coords2 <- as.matrix(read.table(textConnection("
930.3464 7607.313
979.3528 7502.470
865.8662 7484.337
850.5665 7594.947
930.3464 7607.313
")))


poly1 <- st_polygon(list(poly1 = coords1))
poly2 <- st_polygon(list(poly1 = coords2))
pols <- st_sfc(poly1, poly2)
plot(pols)


res <- st_difference(pols[[1]], pols[[2]])
plot(res)
hnagaty
  • 101
  • 1