Drone LiDAR classification: Revealing terrain and flood risk

The most valuable part of a drone LiDAR dataset isn’t always what you see first — it’s what emerges when you start removing the layers.

This flythrough, shared with us by Network Rail asset engineer Ken Durling, shows how a classified point cloud can be used to show the true extent of the terrain — as part of flood-risk assessments to identify low areas in the ground where water might pool or collect.

If the video does not automatically play in HD, click the settings cog and select the highest video quality.

By removing the classified vegetation, structures, and other surface features, subtle terrain features begin to emerge — including drainage paths, depressions, and low-lying areas.

Ken says: "What looks like a uniform landscape from above can reveal complex micro-topography once the upper layers are removed.

"The ability to categorise and analyse LiDAR returns in this way provides a powerful insight tool for infrastructure, environmental monitoring, and flood risk assessment."

The dataset was captured using the DJI M400 and Zenmuse L3 (rented from heliguy™), processed in DJI Terra, and refined in DJI Modify to isolate ground-level returns. The flythrough was created in NUBIGON Inc.

Thanks to Ken for sharing the dataset with us. Our in-house survey team provided Ken with information about L3 data capture parameters to enhance the survey.