I'm trying to calculate metrics for a LAScatalog with the voxel_metrics function. This is the code I used:
library(lidR)
ctg <- readLAScatalog("laspath")
voxel_metrics(ctg, ~mean(Z), 4)
When I run it, I get this message:
Error: Argument is not a LAS object
Can the voxel_metrics function be used with a LAScatalog? (A similar question was asked here, but I don't think the answer applies in this case.)
voxel_metrics does not have a version for LAScatalog. The reason is that the output is a data.frame. Unlike spatial objects like sf, SpatRaster or LAS a data.frame can contain anything. It is thus impossible to let the LAScatalog processing engine automatically handle the logic of a function returning a data.frame.
You could convert it into sf with PointZ geometries and record it into geopackage files. This would looks like this
opt_output_files(ctg) = "/path/file/to/export_{ID}"
opt_chunk_buffer(ctg) = 10
ctg@output_options$drivers$sf$extension = ".gpkg"
apply_vm = function(las)
{
m = voxel_metrics(las, 8, ~mean(Intensity))
m = sf::st_as_sf(m, coords = c("X", "Y", "Z"))
}
res = catalog_map(ctg, apply_vm)
If your dataset if small enough, you can remove opt_output_files(ctg) to return everything in R as one big sf object.
You are free to find another solution that fits your need. This is just a quick solution I can offer right now.
