I have a cloud of points scattered in a 2D Euclidean space. I would like to calculate the area inside the polygon linking the most extreme (=peripheral) points of the cloud. In other words, I would like to estimate the area covered by the cloud in this space. Is there a formula in R? Thanks a lot for any response Julien
Asked
Active
Viewed 4,047 times
1 Answers
16
This is called the convex-hull problem; R built-in chull
function should do the work. To count area, you may use a formula from here.
EDIT: Even better; splancs
package has areapl
function. So the function solving your problem should look like this:
cha<-function(x,y){
chull(x,y)->i
return(areapl(cbind(x[i],y[i])))
}
For instance:
library(splancs);
x<-rnorm(20);rnorm(20)->y;
#Some visualization
i<-chull(x,y);plot(x,y);polygon(x[i],y[i]);
#The area
cha(x,y);

mbq
- 18,510
- 6
- 49
- 72
-
`rgeos::gArea` works with classes created to handle spatial problems (see package `sp`). Might be worth exploring in that direction. – Roman Luštrik Apr 19 '11 at 08:53