Studies
Completed Theses

Completed Theses

COMPLETED BACHELOR THESES

  • Trajectory modeling
    With global navigation satellite systems and their free positioning services, as well as small, low-cost GNSS receivers, it's never been easier to capture and record motion anywhere, anytime. The resulting data volumes can quickly become very large. This makes these records impractical when it comes to storage and evaluation. Approaches to reduce the amount of data while preserving a maximum of spatio-temporal information are required.
    Leaders: Colin Fischer
    Team: Sebastian Leise
    Year: 2016
    Lifespan: 2016

COMPLETED MASTER THESES

  • Optimal assignment of point clouds using deep learning
    The main goal of this master thesis was to register airborne 3D point clouds from different sensor systems. These point clouds are derived from airborne laser scanning (ALS) and dense image matching (DIM) of aerial images. Those point clouds may cover the same surface, but do contain different attributes and characteristics. One major problem when dealing with those two point clouds is vegetation. The laser beam in ALS is able to penetrate vegetation leading to ground and vegetation points in the final point cloud. Since they are derived from aerial images, DIM point clouds only contain the surfaces and thus describes the treetops. This major difference between ALS and DIM causes problems for the registration of point clouds and established algorithms such as the iterative closest points (ICP) algorithm are facing issues when dealing with both point cloud types at the same time, because they assume that similar points are close to each other.
    Leaders: Brenner, Politz
    Team: Stephan Niehaus
    Year: 2019
  • Deep Learning for Flood Relevant Images and Texts from Social Media
    Floods are among Earth's most common and most destructive natural hazards. This work explores the idea of utilizing user-generated information from social media to recognize early signs of flood relevant events. The goal of this work lies in the development and implementation of a Deep Learning solution with the ability to detect the presence of flood relevant events from user-generated images and texts.
    Leaders: Yu Feng, Prof. Brenner
    Team: Sergiy Shebotnov
    Year: 2018
  • Traffic-sign Recognition from Street-level Photos: a Deep Learning Approach
    The scope of this thesis is the recognition of traffic-signs from street-level images. A state-of-the-art deep learning detection algorithm is used, the Single Shot Multi-box detector (SSD) and within the thesis its performance is validated experimentally by examining different training strategies.
    Leaders: Prof. Dr. Bernhard Roth, Co-advisor: Prof. Monika Sester, Supervisor: Stefania Zourlidou
    Team: Qifa Bao
    Year: 2018
    Lifespan: 2018
  • Deep Learning: Deep learning on weather relevant images
    Social media is nowadays an important data source to obtain real time information about the society. As we are now focusing on flooding events, an intelligent interpretation of the user sent photos is helpful for us to find out the occurrence of flooding events and their inundation areas.
    Leaders: Feng
    Year: 2017
  • Automated Calibration of the Camera Orientation of a LiDAR Mobile Mapping System
    The ikg uses a mobile mapping system RIEGL VMX-250 to acquire geo-referenced point clouds and images of the environment. The system includes two laser scanners and four cameras. The coarse camera orientation is estimated by a manual workflow selecting corresponding points in the point cloud and camera images.
    Team: Anita Sadat Khezri
    Year: 2017
    Lifespan: 2017
  • Comparison of Methods for Parking Availability Prediction
    Parking availability prediction is an interesting domain which needs to be explored in order to make the search of a parking place easier, which could in turn save time and energy. Employing pre-existing data (sequential parking data) to develop a model for future prediction is the intent of this thesis. Such a model indirectly reduces the need of static sensors to collect data every five minutes on a daily basis.Different methods, namely, linear regression, Long Short Term Memory-LSTM (a specific recurrent neural network) and Auto regressive integrated moving average (ARIMA) are used in the modeling of data to predict the parking availability.
    Team: Supriya Gurupadaswamy
    Year: 2017
    Lifespan: 2017
  • Extracting Road Network Structure from Heat Maps of GPS Trajectories Using Convolutional Neural Networks
    Road network extraction from GPS trajectories has always been an overemphasized topic for the researches on map improvement. The existence of huge amount of GPS trajectories collected from vehicle movements, cycling and running activities, has encouraged researchers to work on this topic. Besides that, the heat maps (i.e. density maps) of trajectories have been visualized in order to indicate the frequency of use of road segments. With the work represented in this thesis, extraction of road networks from GPS trajectories has been aimed by proposing the use of convolutional neural networks on heat maps.
    Team: Sercan Çakir
    Year: 2017
    Sponsors: DFG Graduiertenkolleg SocialCars
    Lifespan: 2017
  • Gesture-based interaction with virtual 3D environments
    With the availability of increasingly powerful computing technology in the home/leisure sector, a race to develop affordable virtual reality (VR) and augmented reality hardware broke out on the technology market a few years ago, targeting potential markets, in particular, for realistic 3D content presentation (e.g. computer games). The core of this technology is the processing of three-dimensional information in the form of 2D stereo image pairs, which can be consumed via suitable output hardware (glasses/helmets). However, this principle can also be used elsewhere, for example for better exploration of or interaction with 3D spatial data.
    Leaders: Colin Fischer
    Team: Florian Politz
    Year: 2016
    Lifespan: 2016
  • Robust visual navigation for autonomous underwater track vehicles
    Underwater track vehicles, also known as crawler, are universal carrier platforms for many different applications. Crawler having an autonomous navigation would enable the possibility of executing long-term observations without a connection to a base station. This thesis presents approaches that use previous knowledge about the scene that is integrated into motion estimation step by replacing RANSAC with PROSAC to make the motion estimation more robust.
    Leaders: Brenner, Kirchner
    Team: Lewin Probst
    Year: 2015
    Lifespan: 2015
  • Robotic exploration for mapping and change detection
    Autonomous systems and mobile robots become more and more part of our daily life. Examples are cutting the grass in the garden, helping us to get into a parking lot or cleaning the floor. The problems of localization, perception and automatic model building (e.g. maps) are central questions in mobile robotics. How to determine the absolute pose of a robot? What is the best way to explore an a priori unknown environment? Can changes be detected?
    Leaders: Brenner, Paffenholz
    Team: Sebastian Gangl
    Year: 2014
    Lifespan: 2014