How to delete duplicate LiDAR points?

Question

How can I delete duplicate LiDAR points? I incorrectly read the same las twice and saved as a one file resulting in duplicate XY points. How should I delete these duplicate points?

Answer 1

You can delete all duplicate points from a las, laz or ascii file using lasduplicate in LAStools.

Finds and removes all duplicate points from a LAS/LAZ/ASCII file. In the default mode those are xy-duplicate points that have identical x and y coordinates. The first point survives, all subsequent duplicates are removed. It is also possible to keep the lowest points amongst all xy-duplicates via ‘-lowest_z’.

It is also possible to remove only xyz-duplicates points that have all x, y and z coordinates identical via ‘-unique_xyz’.

Also, lasfilterduplicates in the lidR package (documentation p. 51):

lasfilterduplicates(las)

Filter points that appear more than once in the point cloud according to their X Y Z coordinates

Answer 2

If you consider to remove duplicated points on X, Y coordinates (with different Z), after using lasfilterduplicates function from lidR package, you may use lasfilter function.

las <- lasfilter(las, !duplicated(las@data, by = c("X", "Y")))

I think it will retain the first return points (from that duplicated points).

Answer 3

You can also try a tool in WhiteboxTools - LidarRemoveDuplicates. You can run it separately through CMD line or as QGIS 3.4 plugin. Instruction on how to use and install WhiteboxTools in on their webpage

Answer 4

As an update on the answers by @wnursal and @Aaron :

The lidR package function lasfilterduplicates() is deprecated. Instead, you can use the maintained version filter_duplicates(). This can be applied to LAS files and LAS catalog objects.

When loading a catalog or a file, you can already set this within the function, and this maybe works also in your case:

# catalog
ctg <- lidR::readLAScatalog("path/to/folder", filter = "-drop_overlap")

---

More options on filtering the initial data with readLAS(filter = "-help"), not to be confused with later filtering based on Boolean criteria with filter_poi()