Master and Bachelor thesis

Writing a master or bachelor thesis is among the most important aspects of your studies. It is the opportunity to prove your successful studies and your academic abilities and to work together with experienced scientists from your university.

In general, if you are interested in the area of big geospatial data analysis and spatial data science, you might think about starting your thesis project with us. The interest of our chair becomes clear from the research page (Forschung, Research). As soon as you think your interests align with us, feel free to contact us for a meeting.

Here are a few examples of topics we are offering, however, just get in touch and we will find a good topic at the intersection of your interest and our research.

Example Topics

Alternative Path Topology for Point Cloud Analysis

Measured 3D point clouds are one of the most important data sources for autonomous driving, smart cities, and smart manufacturing. However, this data type provides its own challenges. In this thesis, you will work on extracting meaningful topological entities like windows in facades from analyzing families of alternative paths in a graph constructed by joining nearby points in 3D space.

Spatial Quantum Computing

Quantum computing is currently one of the interesting aspects of future computing architectures in that it provides significant speedup to certain clasically hard algorithms. We offer a few theses in which you construct your own quantum algorithms to solve selected spatial problems. We can offer access to a real quantum computer (IBM Q) through collaboration with our partners from the CODE research center at UniBW.

Bot Rejection from Precise Spatial Knowledge

Social media comprises one of the hot topics in big gespatial data analysis. But it is very hard to use as it is very noisy and in parts actively influenced by campaigns and bots. In this thesis project, you will extend an existing bot rejection scheme based on spatial computing and investigate the effect of bot and noise removal on a selected set of machine learning tasks involving social media data.

Indoor Mapping made cheap

Indoor Navigation and more generally acquiring spatial data in buildings is a challenging yet very valuable task. In this engineering thesis, you will build a simple indoor mapping system based on modern sensors and algorithms and analyze the impact of data fusion from cheap and expensive sensors.

Social Media Search and Retriveal

Internet search engines are widely used and accepted for organizing the knowledge available from the world wide web. In this thesis, you will extend our existing social media search engine with a set of aspects related to spatial influence and toponym mining. In this way, users are enabled to use keyword search together with spatial search in a flexible joint framework. This work is related to spatial analysis, spatial autocorrelation, information retrieval and machine learning.

Helena Fusion

Helena fusion is a project in which we design a novel high-level programming langauge enabling the automation of spatial machine learning flows and cross-sensor data fusion. Based on preliminary work in this area, we want to automate the process of spatial information mining and integrate a layer of data acquisition and coregistration. In this thesis, you contribute to this prestigious work and help us exploit cluster computing architectures for remote sensing image analysis in a systematic way.

Compression for Climate Land Cover Classification (ESA CCI)

Land cover classification is one of the more traditional areas of remote sensing. One of the application domains is climate change research. The ESA provides climate-relevant land cover data from an exceptionally interesting project Unfortunately, the data footprint of such global landcover mapping projects is huge and so is the amount of energy wasted in communicating and analyzing such data. In this thesis, we will discuss methods for early compression of data in projects related to climate-related land cover classification.

© 2020 M. Werner