I have been using a Reigl TLS scanner on vegetation, and I want to use the intensity values to help classify which points are vegetation, and which are 'other' for instance branch wood or stem wood in trees. Several attempts at coding this in Matlab have not only proved inconsistent results from scan to scan, but also highlight that there are inconsistencies in the intensity values when the scans were obtained in different solar light levels.
Can anyone recommend a methodology to help me consistently classify vegetation points from 'other', using LiDAR intensity values created via a TLS? Thanks in advance.