How to get las Zref values after normalization in lidR?

Question

I understand that normalize_height() in lidR creates a Zref attribute so that it can be unnormalized. While I can print Z for the normalized file, I cannot print Zref. Subsequently, when I apply find_trees() Z values are carried over but of course Zref again returns NULL. My goal is to have tree tops with height and original elevation (so that it can be displayed in software that allows it to be seen in 3D along with imagery) and I was thinking this would be more efficient than extracting values from the raster DEM.

myLas <- readLAS('C:/ITD/normtest.las')
ttops <- find_trees(myLas, lmf(ws = 5))
ttops$Z     # this prints a series of values for tree heights
ttops$Zref  # this results in NULL

Answer 1

The use case you are describing is not natively covered. You can normalize a point cloud. In this case the output is a LAS object with an attribute Zref that can be unnormalized easily

library(lidR)
LASfile <- system.file("extdata", "Topography.laz", package="lidR")
las <- readLAS(LASfile)
nlas <- normalize_height(las, tin())
head(nlas$Zref)
#> [1] 806.5340 806.5635 806.0248 809.6303 809.3880 809.4847

However the output of find_trees is SpatialPointsDataFrame with two attributes: an ID and the Z elevations.

ttops <- find_trees(nlas, lmf(ws = 5))
ttops
#> class       : SpatialPointsDataFrame 
#> features    : 2047 
#> extent      : 273357.3, 273642.8, 5274357, 5274643  (xmin, xmax, ymin, ymax)
#> crs         : +proj=tmerc +lat_0=0 +lon_0=-70.5 +k=0.9999 +x_0=304800 +y_0=0 +ellps=GRS80 +units=m +no_defs 
#> variables   : 2
#> names       : treeID,        Z 
#> min values  :      1,  2.00275 
#> max values  :   2047, 20.97725

You could retrieve the original points by matching coordinates but I really think that the simplest way is to add the DTM elevation

dtm <- grid_terrain(las, 0.5, tin())
Zdtm <- dtm[ttops]
ttops$Z <- ttops$Z + Zdtm

plot(las) |> add_treetops3d(ttops)

or maybe even simpler/faster using a CHM

chm <- grid_canopy(las, 0.5, p2r())
ttops$Z <- chm[ttops]

plot(las) |> add_treetops3d(ttops)