Knowledge Base & Support

Ground (UAV surface filtering) Classification Model


This is a classification model that separates the ground plane from the remaining data and applies filtering on flat surfaces to better average the result from UAV data, placing points filtered out into the noise class.

Low ground and high ground together represent the ground plane detected. Low ground represents the lowest points detected as ground, similar to the representation from a cloth-draping algorithm.

Ground detection is often a core requirement for DEM generation and other bare earth terrain management and mapping purposes.

Optimal Scene

Large scale aerial coverage of regions

Optimal Scanner Type

UAV laser scan point cloud data

Optimal Average Point Spacing

25 mm or closer


Class List

Class NameLAS/Z code
low ground2
high ground65
non ground64