Profile picture of Felix Dahle

Felix Dahle

I joined the group in March 2021 and I am supervised by Bert Wouters and Roderik Lindenbergh. My research interests are in the fields of Geo-informatics, Computer Vision, Photo­grammetry and Machine Learning. I obtained a Master degree in Geomatics at TU Delft, during which I worked as a developer for Readar.

Current Research

My PhD research is focused on historical changes of Antarctic glaciers from photogrammetry: I am studying historical aerial imagery from the Antarctic continent in order to derive historical DEMs. These models will be compared with present day elevation data to obtain a detailed picture of elevation and mass changes over the past 50 years. In this way the long-term response of the Antarctic Peninsula to changing climate conditions and ice shelf thinning can be studied.

Publications

  • Polar perspectives: a deep dive into geo-referencing historical Antarctic photos

    Felix Dahle, Bert Wouters, Roderik Lindenbergh
    We developed a workflow to geo-reference historical aerial imagery with a use-case on the Antarctic Peninsula. Using modern algorithms for tie-point matching (LightGlue), we can find similar points on the historic images and modern Sentinel-2 imagery and use these points for geo-referencing. A special focus in this paper is the tie-point matching, which was adapted to work with historical and large imagery.
    International Journal of Digital Earth

  • Revisiting the Past: A comparative study for semantic segmentation of historical images of Adelaide Island using U-nets

    Felix Dahle, Bert Wouters, Roderik Lindenbergh
    Building on the previous article we are extening the semantic segmentation of historical photographs of the Antarcitca using U-nets. We compare multiple hyper-parameters and augmentation methods to identify the optimal settings for a segmentation of Adelaide Island.
    ISPRS Open Journal of Photogrammetry and Remote Sensing

  • Semantic segmentation of historical photographs of the Antarctic Peninsula

    Felix Dahle, Julian Tanke, Bert Wouters, Roderik Lindenbergh
    Semantic segmentation is applied to historical photographs of the Antarctica using a U-net based machine learning approach. Our results show that our method can handle very challenging images even after being trained with only a low number of training data and catch the general semantic meaning of a scene.
    XXIVth ISPRS Congress 2022 (Nice, France)

  • Automatic change detection of digital maps using aerial images and point clouds

    Felix Dahle, Ken Aroyo Ohori, Giorgio Agugiaro, Sven Briels
    In this work we are showing that random-forest based approaches can be used to detect changes between two time steps using 2.5D data and to transfer these insights to a digital map. Results show that this methodology can recognize a substantial amount of changes and can support - and speed up - the manual updating process
    XXIVth ISPRS Congress 2021 (Virtual)

  • CITYJSON + WEB = NINJA

    Stelios Vitalis, Anna Labetski, Freek Boersma, Felix Dahle, Xiaoa Li, Ken Arroyo Ohori, Hugo Ledoux, Jantien Stoter
    3D city models would greatly benefit from a proper web-based solution to visualise and manage them. In order to provide an implementation of a web application for CityJSON data, that can be used as a reference for other applications, we developed ninja. It is a web application that allows the user to easily load and investigate a CityJSON model through a web browser.
    XXIVth ISPRS Congress 2020 (Virtual)

  • Europe's socio-economic disparities reflected in settlement patterns derived from satellite data

    Hannes Taubenböck, Felix Dahle, Christian Geiß, Michael Wurm
    In this work we detect urban nodes as anchor points of urban densification. We identify a network of cities when conjugation lines between these urban nodes feature high settlement density. Further, we map regions around connected urban nodes with high settlement densities. We assume that these identified regions belonging to a network of cities express beneficial economic development.
    JURSE 2019 (Vannes, France)
    Paper | DOI | Bibtex