rcaiman (version 2.0.1)

extract_feature: Extract feature

Description

Extract a numeric or logical summary from segmented raster regions using a user-defined reducer, returning one value per segment as a raster map or a named vector.

Usage

extract_feature(
  r,
  segmentation,
  fun = mean,
  return = "raster",
  ignore_label_0 = TRUE
)

Value

If return = "raster", a terra::SpatRaster where each pixel holds its segment’s feature value. If return = "vector", a named numeric (or logical) vector with one value per segment.

Arguments

r

numeric terra::SpatRaster with one layer.

segmentation

single-layer terra::SpatRaster. Segmentation map of r, typically created with functions such as sky_grid_segmentation(), ring_segmentation() or sector_segmentation(), but any raster with integer segment labels is accepted.

fun

function taking a numeric/logical vector and returning a single numeric or logical value (default mean).

return

character of length one. Either "raster" (default) or "vector", controlling whether to return a map with per-segment values or a named vector (one value per segment).

ignore_label_0

logical of length one. If TRUE, the segment labeled 0 is ignored.

Details

Segments labeled 0 can be ignored via ignore_label_0 = TRUE. The function in fun must return a single numeric or logical value for any input vector (e.g., mean, median, or a custom reducer).

Examples

r <- read_caim()
z <- zenith_image(ncol(r),lens())
a <- azimuth_image(z)
g <- sky_grid_segmentation(z, a, 10)
print(extract_feature(r$Blue, g, return = "vector"))
# plot(extract_feature(r$Blue, g, return = "raster"))