Student information

Msc thesis subject: Automatic feature detection in high resolution DEMs and UAV-LiDAR Point clouds for Archaeology

The possibility to acquire very dense LiDAR data from UAVs is opening up a new dimension in archaeological research. Larger areas can be scanned in short time and the level of detail is showing features that were left undiscovered before.

The main goal of the thesis is to develop (or evaluate existing) algorithms to automatically determine archaeological features in a LiDAR derived DEM, or directly in the point cloud.

In the last years we performed a few flight campaigns with the RiCOPTER of terrain with archaeological features. (Groenlo: Spanish Defence lines, Hurtgenwald: 2nd world war bunkers, trenches and bomb craters, Heede: remains of a medieval castle). The features can have different shapes, so the method should be able to handle linear, circular or odd-shaped features.

All areas were covered by forest, which means that filtering out the vegetation is an important step in the processing. But how do the choices in the vegetation filtering / DEM generation influence the outcomes of the feature detection?

Objectives

  • Implement or develop a method to automatically detect archaeological features in LiDAR derived DEMs or point clouds.
  • Evaluate how choices in vegetation filtering and DEM generation influence your feature detection.

Requirements

  • Affinity with laser scanning

Theme(s): Sensing & measuring, Modelling & visualisation