Output rasters have offset origin

[zoeschindler]

Hello, I have an issue with the rasterize_canopy() function. The output raster is not centered at (0,0), this causes the raster to not perfectly overlap with some of my other rasters created in lidR and causes me further issues. The issue starts at resolutions lower than 0.5 meters, but I need a raster with 0.1 meters resolution. I noticed this also happens for example when creating rasters with pixel_metrics().

Here is an example:

# load data
LASfile <- system.file("extdata", "Topography.laz", package="lidR")
las = readLAS(LASfile)

# calculate dsm
dsm <- rasterize_canopy(las, res = 0.5)
dsm@ptr$origin
#> [1] 0 0

# calculate dsm
dsm <- rasterize_canopy(las, res = 0.1)
dsm@ptr$origin
#> [1] 1.117587e-08 3.438443e-06

[Jean-Romain]

Minimal reproducible example

library(lidR)
LASfile <- system.file("extdata", "Topography.laz", package="lidR")
las = readLAS(LASfile)
u = lidR:::raster_layout(las, 0.1)
u = lidR:::raster_materialize(u)
u@ptr$origin
#> [1] 1.117587e-08 3.438443e-06

[Jean-Romain]

More minimal reproducible example. The issue comes from terra. Reported in https://github.com/rspatial/terra/issues/844

options(digits = 14)
b <- c(273357.1, 273642.9, 5274357.1, 5274642.9)
b <- terra::ext(b)
r <- terra::rast(b, resolution = 0.1)
r
#> class       : SpatRaster 
#> dimensions  : 2858, 2858, 1  (nrow, ncol, nlyr)
#> resolution  : 0.099999999999996, 0.099999999999935  (x, y)
#> extent      : 273357.1, 273642.9, 5274357.1, 5274642.9  (xmin, xmax, ymin, ymax)
#> coord. ref. :
r@ptr$origin
#> [1] 1.1117663234472e-08 3.4375116229057e-06

^{Created on 2022-10-12 by the reprex package (v2.0.1)}

[zoeschindler]

Thank you very much for your efforts!

[rhijmans]

@zoeschindler: there is not much that can be done about this imprecision. The "further issues" you mention can be addressed, as small differences like this are supposed to be ignored. But I do not know what the further issues are; so please report these, if you can.

[zoeschindler]

@rhijmans I had issues when using resample(). I resampled to a higher resolution and got too many raster cells in my new raster, as the cells of the fine raster slightly overlapped the cells of the rough raster I resampled to. I mentioned it in https://github.com/rspatial/terra/issues/847. I am not sure how much of this problem is attributed to the slight offset or the issue with the function though.