Fine-Grained Point Cloud Semantic Segmentation of Complex Railway Bridge Scenes from UAVs Using Improved DGCNN

Figure 3

Schematic diagram of 7 segmentation elements. (a) Element annotation in real-world scenes; (b) element annotation in point cloud scenes. The different elements are defined as follows: cable refers to all visible cables; mast refers to all kinds of masts; rail refers to all railway tracks; track bed represents the collection of structures under the tracks; protective wall refers to the protective walls of the bridge section; guardrail refers to the guardrails on both sides of the bridge; cluster refers to the collection of points in the point cloud that do not have any semantic meaning.

(a)

(b)