search

r-lidar / lidR

Airborne LiDAR data manipulation and visualisation for forestry application
https://CRAN.R-project.org/package=lidR
GNU General Public License v3.0
606 stars 132 forks source link

Time taken by lasground and grid_terrain commands? #74

Closed ahmadshabbir2468 closed 7 years ago

ahmadshabbir2468 commented 7 years ago

I am beginner to pointcloud and R. I setup R in ubuntu and install LidR package. i want to classify Tree from las file. I am using steps given in "Tree-segmentation-from-A-to-Z". when i run lasground it takes long time and nothing happens same in the grid_terrain case. Anyone here tell me what is the reason behind it?

Jean-Romain commented 7 years ago

Hi,

What do you mean my nothing happens ? The script crashed ? The computation is running over and over ? The computation ended but you don't know what to do with the result ? Could you show your code ? Are you asking why it takes a long time ? If yes how many points do you have ? Please provide more information if you ask for help.

ahmadshabbir2468 commented 7 years ago 
library(lidR)
las = readLAS("test.las")
plot(las)
lasground(las, MaxWinSize = 10, InitDist = 0.05, CellSize = 7)
dtm = grid_terrain(las, method = "kriging", k = 8)
plot3d(dtm)

After readLAS, lasground commands runs and is not shifted to nextline of grid_terrain. I run lasground 2 hours ago and at this time they are at samepoint not shifted to nextline grid_terrain. :/

Jean-Romain commented 7 years ago

I don't know what you mean by "shifted to nextline". If you mean that the dataset is not normalized with the ground at 0 this is normal. lasground classifies the points as ground and not ground. lasnormalize enable to subtract the terrain. Two hours is long. I know that the function is not blazing fast but you may have a lot of points... Also the parameters are maybe not appropriated to your terrain. You're better to used the classification made by the data provider if available.

You should try your script with a small subset of your dataset before to apply it to big data. At least to test if it does what you expect it to do.

library(lidR)

LASfile <- system.file("extdata", "Topography.laz", package="lidR")
las = readLAS(LASfile, Classification = FALSE)
plot(las)

lasground(las, MaxWinSize = 30, Slope = 2, InitDist = 0.01, CellSize = 7)
plot(las, color = "Classification")

dtm = grid_terrain(las, method = "kriging", k = 8)
plot3d(dtm)

las = lasnormalize(las, dtm)
plot(las)

Also notice that sometime the wiki pages are up-to-date with the development version but not with the released one.