I would like to process an unclassified point cloud data created by photogrammetric technique. On the area of interest of an urban area can be found family houses, vegetation, also it's a relative flat area with detailed objects, many spectral properties.
The source datas are scanned (!) aerial images, I created manual GCPs with real xyz coordinates to create an unclassified point cloud in LAS format or in simple ascii format.
I would like to use the plugin CSF / or another plugin to achieve two layers, of course ground points and non-ground points dataset.
I tried the CSF (very good method!!) but I have the main question, what's another way to achieve my target without many classified errors? So my clipped area contains ~300000 points with real xyz coordinates, the resolution is low and can be found holes, too.
Thanks a lot for the answers.