I want to create a ground model from a *.las file. The LAStools are available in QGIS but limited in use (generation of black diagonal in result file), so I can't use this.
In the GRASS tools I can find the tools v.lidar.edgedetection, v.lidar.growing and v.lidar.correction, to classify the las data. Unfortunately, v.lidar.edgedetection gives me a python error. So this isn't working either.
How could I proceed?
It is possible to generate a DEM from a las file within QGIS using Fusion.
Use this tutorial to enable FUSION's tools in QGIS (works up to QGIS 2.18 version*). It is written for LAStools, but it is straightforward adapting it.
*As from QGIS version 3, one can install the plugin (besides the core software) and work from there. See: Unable to install FUSION in QGIS 3.2?.
Once you manage to activate Fusion, you will need the GroundFilter tool for classifying ground points (if the point cloud is not yet classified), and then, GridSurfaceCreate to generate the DEM.
If you don't need to use QGIS, I suggest the following approach, which classifies ground points with the MCC algorithm.
I am using the following procedure in SAGA:
pc_processing_with_saga (Wichmann et al. (2012)) deprecated link,
new link!
If you still want to use the DEM in other programs (like QGIS), export it as GeoTIFF from SAGA when reaching Close Gaps step (I use already classified las files).
In SAGA it works like a charm and it is fast.
pktools, when built with liblas support, provides functionality to generate DEMs from LAS point clouds. The toolkit provides a program called pklas2img, which rasterises a LAS/LAZ point cloud using filtering/compositing options. For instance, the following commands would generate a DEM:
pklas2img -i input.las -o output_dem_min.tif -comp min -n z -dx 2 -dy 2 -ot Float32
Where:
comp is the compositing rule for multiple points in a cell, in this example we use min, which retains the pulse with the minimum height for each grid cell. The documentation provides the full list of compositing options.dxand dy are the output resolutionsAfter this step, it is advisable to filter the DEM, which can be done using pkfilterdem as follows:
pkfilterdem -i output_dem_min.tif -o output_dtm.tif -f promorph -dim 3 -dim 17
Where:
Note, this functionality is also available from the Processing Toolbox in QGIS using pktools version 2.6.6 (see below):
Install PDAL library (with OSGeo4W) and look at this example https://pdal.io/workshop/exercises/analysis/dtm/dtm.html
The lidR package is an option if you are comfortable writing some simple R code. The method grid_terrain will interpolate a digital elevation model (ground surface) from a classified point cloud. Several algorithms are available to do this including tin (triangulation) and knnidw (k nearest neighbor inverse distance weighting). If your point cloud has no classification, the package also offers a ground classification function.
