Segmentation algorithms fall into broadly four groups:

1.Pattern based techniques – Cluster analysis

2.Edge detection techniques

3.Graph based techniques

4.Region growing techniques

Using geometric characteristics

Typical geometric characteristics of a segment include position/location, orientation, roughness/texture, perimeter length, area, elongation/eccentricity and compactness

Using radiometric characteristics

Every point in a point cloud is associated with at least one radiometric characteristic, i.e., the reflectance strength of the returned pulse. Some ALS systems also capture an image simultaneous with the pulse measurement, thus allowing an RGB triplet to be associated with each point

Using topological characteristics

The topology, i.e., the spatial relationship between segments, can be described by the adjacency and connectivity of segments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.