Merge overlaid las files

[seanreilly66]

Hi everyone,

I have two registered las files that I am attempting to merge. Is there a recommended way to do this within the lidR package? Currently, I have come up with the following but this method is slow for pairs of large tls las files (>3 gb):

library(lidR)

las1 <- LAS(data = data.frame(
  X = runif(100, 1, 10),
  Y = runif(100, 1, 10),
  Z = runif(100, 1, 10)
))

las2 <- LAS(data = data.frame(
  X = runif(100, 1, 10),
  Y = runif(100, 1, 10),
  Z = runif(100, 20, 30)
))

merge_df <- rbind(las1@data, las2@data)

merge_las <- LAS(merge_df)

Thank you! I really appreciate any suggestions.

Sean

[Jean-Romain]

You can use the catalog_retile being clever on how to make chunks

ctg = readLAScatalog(...)
opt_chunk_size(ctg) = 2000 #large enough to encompass your tiles
opt_chunk_alignment(ctg) = c(...) # may be required to make only one chunk
plot(ctg, chunk = T) # to check if there is only one chunk
new_ctg = catalog_retile(ctg)

This should use very little memory. Yet the output file will be large.

You can also have a look to my new package lasR. There is merge file section in the tutorial https://r-lidar.github.io/lasR/articles/lasR2.html#write-las

[seanreilly66]

Thank you for the reply and everything you've done with lidR. This code worked great and I will definitely give lasR a try.

[seanreilly66]

Is there a way to feed the output from this (e.g., new_ctg) into cloud_metrics without first re-reading from disk? I apologize if there is an answer to this already elsewhere on this page but I have not been able to find it.

[Jean-Romain]

Nope, cloud_metric() calls a user defined R function to your point cloud and thus requires your point cloud to be loaded in memory in a data.frame. catalog_retile() is full c++ and is capable of retiling without loading the point cloud in R memory. Both a very incompatible.