Fast raster extraction at points (buffered)

Extracts summary statistics from a SpatRaster at point locations, optionally using buffered extraction with custom kernel windows.

Usage

fastextract(x, y, d = 0, w = "circle", fun = "mean", na.rm = TRUE)

Arguments

x

SpatRaster. Input raster (single- or multi-layer).

y

SpatVector. Points or polygons.

d

numeric or numeric vector. Buffer radius/radii in map units.

w

character. Window type for the buffer kernel when d > 0 (currently passed through to terra; e.g., "circle", "rectangle").

fun

character or function. Summary function: "mean", "sum", "min", "max", "sd", or "median"; or a user function.

na.rm

logical. Whether to remove NAs when computing summaries.

Value

A data.frame of extracted values. When d has multiple values, rows are stacked by scale with a scale_m column indicating the radius.

Details

• If d > 0, a buffer of radius d (map units) is created around each point and the summary is computed over raster cells intersecting the buffer.

• If d == 0, values are taken at the point locations (no buffering).

• If y is a polygon layer, the summary is computed over polygon areas.

Examples

r <- terra::rast(nrows = 10, ncols = 10, xmin = 0, xmax = 100, ymin = 0, ymax = 100)
terra::values(r) <- seq_len(terra::ncell(r))

pts <- terra::vect(
  matrix(c(10, 10,
           50, 50), ncol = 2, byrow = TRUE),
  type = "points",
  crs  = terra::crs(r)
)

# Mean over a 20-unit circular neighborhood around each point
res <- fastextract(r, pts, d = 20, w = "circle", fun = "mean")
head(res)
#>    lyr.1
#> 1 83.125
#> 2 50.500