Labelling Biodiversity in 3D Forests (MSc topic)

Background: With the introduction of lidar (Light Detection and Ranging) sensors on drones and tablets, the amount of 3D data available for research and education is set to increase drastically over coming years. While collecting 3D data is no longer a barrier, existing challenges around efficiently adding semantics to 3D data become more evident. Most commonly, the segmentation of 3D data handled manually using software such as CloudCompare (Fig. 1) which can be time-consuming and tiresome. Alternatively, an automatic approach based on machine learning algorithms could be used. However, these methods require training data. Training data is the initial dataset used to train machine learning algorithms. Models create and refine their rules using this data. To improve the quality of training data for 3D point cloud analyses, and to improve overall labelling precision as well as the workflow for doing so, this project aiming to develop a VR (Virtual Reality) labelling application.
By using computer tools such as CloudCompare, DendroCloud, and the Unity engine and conducting a user-study you will answer related questions such as:

• How does forest labeling in VR perform compared to traditional segmentation methods?
• What accuracy can be achieved by labelling forest in VR?
• What are the limitations of labelling forest in VR?

Wanted: We are looking for a student with solid knowledge of forestry, (eg. obtained via courses such as 701-0303-00 Waldvegetation und Waldstandort, or the Systemvertiefung Wald & Landschaft) and a strong interest in forest digitisation.
No previous experience with VR tools is required, but an interest to learn is essential.

The project has a flexible starting date, with preference given to interested candidates who would like to start as soon as possible.

You will get to:

• Work on an interdisciplinary, cutting-edge topic, with a lot of room for creativity;
• Expand your network discussing your work with experts from the intersecting fields of forest sciences, geomatics and computer sciences;
• Be a co-author on a publication resulting from this work;
• Be part of a motivated, fun, and energetic team of scientists.

Supervisor: C. Fol, Dr. A. Murtiyoso, Prof. Dr. V. Griess

If the idea of participating in cutting edge interdisciplinary research excites you, please contact cyprien.fol(at)usys.ethz.ch. The FORM team is looking forward to hearing from you!