CODEWITHSUNDEEP
rgispolygonrasterspatial
Muesgen
Muesgen

Reputation: 162

How to check if raster and polygon are overlapping?

i have a raster and i want to check if the cells with values are overlapping my area of Interest, which is a spatialpolygon file.

Is there a fast way to do this?

My raster look like this:

enter image description here

what i tried:

r <- raster("C:/user/test.jp2")
studyarea <- readOGR("C:/user/boundary.shp")
dummi <- as(extent(r), "SpatialPolygons")
if (gContainsProperly(studyarea, dummi)) {
  print("Raster extent is fully within the Area of interest")
} 
else if (gIntersects(studyarea, dummi)) {
  print("Raster extent is not fully within the Area of interest")
} else {
  print("Raster extent is fully outside the Area of interest")
}

as(extent(r), "SpatialPolygons") creates a polygon within the values and also the NAs. But i only want to check if the value part and the studyarea are overlapping

Data example for the raster and studyarea:

matrix <- matrix(c(NA,NA,NA,1,NA,NA,2,1,NA,1,3,4,2,3,4,1), nrow=4)
r <- raster(matrix)
extent(r) <- c(399960, 509760, 5290200, 5400000)
crs(r) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
writeRaster(r, paste(getwd(),"filename.jp2"))
                 
matrix <- matrix(c(1,1,NA,NA,1,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA), nrow=4)
rpoly<- raster(matrix)
extent(rpoly) <- c(399960, 509760, 5290200, 5400000)
crs(rpoly) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
studyarea <- rasterToPolygons(rpoly)

thanks

Upvotes: 2

Views: 2901

Answers (2)

Muesgen
Muesgen

Reputation: 162

I found a quicker solution. It is possible to crop the raster by the polygon and than mask the raster by the polygon. If there are no raster Values in the area of the Polygon. The Raster will have min values = 0 and max value = 0.

Create Data

matrix <- matrix(c(NA,NA,NA,1,NA,NA,2,1,NA,1,3,4,2,3,4,1), nrow=4)
r <- raster(matrix)
extent(r) <- c(399960, 509760, 5290200, 5400000)
crs(r) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
writeRaster(r, paste(getwd(),"filename.jp2"))
                 
matrix <- matrix(c(1,1,NA,NA,1,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA), nrow=4)
rpoly<- raster(matrix)
extent(rpoly) <- c(399960, 509760, 5290200, 5400000)
crs(rpoly) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
studyarea <- rasterToPolygons(rpoly)

Solution

# Reproject. Make sure both have the same projection
studyarea <-  spTransform(studyarea, projection(r))
  
# Crop and clip raster with study area polygon
r_crop <- crop(r, studyarea)
r_mask <- mask(r_crop, studyarea, updatevalue = NA, updateNA= TRUE)

if (r_mask@data@min == 0 && r_mask@data@max == 0){
  print("raster and studyarea do not overlap))
  }
else {
  print("Raster and studyarea do overlap")
}

Upvotes: 2

user9256562
user9256562

Reputation:

So here is an approach using the data example.

The Idea is to convert the raster cells with values to a single polygon, and comparing this polygon with your study areas polygon.

sf provides a wide variety of predicates to compare geometries.

matrix <- matrix(c(NA,NA,NA,1,NA,NA,2,1,NA,1,3,4,2,3,4,1), nrow=4)
r <- raster::raster(matrix)
raster::extent(r) <- c(399960, 509760, 5290200, 5400000)
raster::crs(r) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
r <- sf::st_as_sf(raster::rasterToPolygons(r)) # This will create a sfc polygon dataframe from your cells with values
r <- sf::st_union(r) # This will combine all cells with values to a single polygon

matrix <- matrix(c(1,1,NA,NA,1,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA,NA), nrow=4)
rpoly<- raster::raster(matrix)
raster::extent(rpoly) <- c(399960, 509760, 5290200, 5400000)
raster::crs(rpoly) <- "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs" 
studyarea <- sf::st_as_sf(raster::rasterToPolygons(rpoly)) # This will create a sfc polygon dataframe from your cells with values 
studyarea <- sf::st_union(studyarea) # This will combine all cells with values to a single polygon

sf::st_contains(r, studyarea, sparse = FALSE) #  This checks if studyarea is within r, see ?sf::st_contains for other predicates you might want (touches, crosses, etc.)

Upvotes: 3

Related Questions