Options for making a DEM of single-return point cloud?

Question

I have a project site surveyed with single return photogrammetry so, all points are unclassified. It was not a problem to create a DSM of this data, but I also need a DTM.

What are my options? Will I be able to use lasground from LAStools if all the points are only reflections from the surface? There are areas of bare ground between trees. Is that enough?

Answer 1

I am not as familiar with LasTools but you can try using a variety of tools in WhiteboxTools (https://www.whiteboxgeo.com/). Even though they mention "LiDAR" in the toolname/user manual, it may also be applicable to photogrammetric point clouds. First you can use the LidarGroundPointFilter (https://www.whiteboxgeo.com/manual/wbt_book/available_tools/lidar_tools.html#LidarGroundPointFilter), which performs a slope-based classification, or filtering (i.e. removal), of non-ground points within a point-cloud. This would help to establish ground points.

You could also try the RemoveOffTerrainObject (https://www.whiteboxgeo.com/manual/wbt_book/available_tools/geomorphometric_analysis.html#RemoveOffTerrainObjects) tool on your DSM file(s).

You can check out this paper thats shows how the RemoveOffTerrainObject method can be effective at removing vegetation and buildings from DSMs, albeit it uses LiDAR, but it would be equally well suited to photogrammetrically derived DSMs as well.

(https://www.researchgate.net/publication/323003064_A_new_method_for_the_removal_of_off-terrain_objects_from_LiDAR-derived_raster_surface_models)

This is an example image from the paper and a before/after image of the RemoveOffTerrainObject tool.