Hello,
as mentioned in a few footnotes but not yet properly announced. There is a new addition to LAStools that can help you get rid of isolated noise points in your LAS / LAZ data as they occasionally happen when the laser beam hits a few cloud flurries, some birds, or a swarm of bees (-: or when a belated reflection produces a return well below the ground.
Here a typical use case and attached is the result. The points that are now pink/lilac were reclassified into classification code 7 which is the default mode. There is also an option to remove the points from the file.
lasnoise -i tiles_raw\tile_654000_1614000.laz ^
-step 2 -isolated 5 ^
-odix _denoised -olaz
The lasnoise tool flags or removes noise points in LAS/LAZ/BIN/ASCII files. The tool looks for isolated points according to certain criteria that can be modified via '-step 3' and '-isolated 3' as needed. The default for step is 4 and for isolated is 5.
The tool tries to find points that have only few other points in their surrounding 3 by 3 by 3 grid of cells with the cell the respective point falls into being in the center. The size of each of the 27 cells is set with the '-step 5' parameter. The maximal number of few other points in the 3 by 3 by 3 grid still designating a point as isolated is set with '-isolated 6'.
By default the noise points are given the classification code 7 (low or high noise). Using the '-remove_noise' flag will instead remove them from the output file. Alternatively with the '-classify_as 31' switch a different classification code can be selected.
It is also important to tell the tool whether the horizontal and vertical units are meters (which is assumed by default) or '-feet' or '-elevation_feet'. Should the LAS file contain proper projection VLRs then there is no need to specify this explicitly. Internally the step size in xy or in z direction will be multiplied with 3 if the coordinates are found or are specified to be in feet. The size of the grid cells needs to be changed with '-step 0.0001' or similar if the input is in geographic longlat representation.
example usage:
>> lasnoise -i lidar.las ^
-remove_noise ^
-o lidar_without_noise.laz
removes all points that have only 5 or fewer other points in their surrounding 3 by 3 by 3 grid (with the respective point in the center cell) where each cell is 4 by 4 by 4 meters in size.
>> lasnoise -i tiles_raw\*.laz ^
-classify_as 31 ^
-step 2 - isolated 2 ^
-odir tiles_denoised\ -olaz ^
-cores 3
running on whole folder of files on 3 cores where each cell being only 2 by 2 by 2 meters in size and where points that have only two or fewer other points in the 27 cell neighborhood of their surrounding 3 by 3 by 3 grid are considered noise. here points are classified to class 31 instead of being removed.