Cross-section based dense cloud editing / classification

[bortoman]

While being an amazing Software for photogrammetry, the biggest challenge in PhotoScan is cleaning up a dense cloud and classify it to get a proper DTM.
While the automatic classification algorithm is doing a good job, there is always some (often a lot) manual work to do.

I had the chance to work with Global Mapper for a while, and one feature that was very helpful was the ability to view a cross-section of a point cloud and the ability to select points from there to be modified and classified. You can see how it works here: https://www.youtube.com/watch?v=v8VM2vkw43g&feature=youtu.be&t=3420

Long story short: you define a cross section path and a range (example +-0.5m) and all the points lying in this slice are viewed in a cross-section window. From there you can select points and edit them.

In the attached screenshot i generated a length-profile with a cross-section every 1m...

[JMR]

+1

[Thibaud Capra]

Being able to train an algorithm using a few cross sections to further improve an upcoming ground extraction would be awesome too (I know I'm asking for a lot!)

+1 though.