My research deals with the computer science aspects of managing, storing and processing geospatial data. My main research topic is revision management and change analysis of spatial vector data models: I have developed an operation-based revision control model that makes the capabilities familiar from software version control systems — branching, merging, conflict resolution and the traceability of the modification history — available for geographic features, while taking their specific properties, including topological relationships, into account. I investigate the representation of spatial data at a higher level of abstraction using topological data structures, and how such representations can be embedded into processing environments built on distributed systems. My research also covers the photogrammetric and image processing workflows of aerial imagery captured by drones, including the evaluation of multispectral imagery and the application of computer vision methods in a geospatial context. Throughout my work I rely on free and open source geospatial software, and I publish my own results as open source tools as well.

My research addresses the automated processing of remotely sensed data, primarily point clouds acquired by airborne and terrestrial laser scanning (LiDAR) and multispectral satellite imagery. I develop methods for the automatic recognition and classification of objects over large areas, such as the detection of railway infrastructure (rails, overhead cables) and its defects from mobile laser scanning data, as well as the segmentation of buildings and woody vegetation in urban environments. A central topic of my work is change detection based on multi-temporal acquisitions, which makes it possible to track the transformation of built-up areas and vegetation over time at the level of individual objects. A distinct research direction is the detection of plastic waste and illegal waste deposits in riverine and near-water environments from multispectral satellite imagery using machine learning methods, validated against field surveys and citizen science observations. Since the size of nationwide datasets exceeds the limits of conventional processing, a substantial part of my research concerns the application and comparative performance evaluation of big data technologies and distributed, cloud-based computing architectures.

My research centres on the static analysis of source code, primarily in order to support program comprehension. We develop methods and tools for analysing large, often multi-million-line legacy C/C++ and C# systems, and for making the extracted information accessible to developers through visualisations, dependency and component graphs, and interactive interfaces. A key research direction of mine is incremental source code analysis, which reprocesses only the affected parts of a codebase after a change instead of performing a full re-analysis, thereby substantially reducing the resource demand of the process; within this line of work I also address the incremental parsing of build systems. I further investigate how comprehension features can be delivered into the source code editors used by developers on a daily basis via standard protocols such as the Language Server Protocol. My research extends to project management as well: by analysing the development history stored in version control systems, I devise methods for measuring developer expertise and the distribution of knowledge over a codebase.

My research focuses on tool support for the software development process and on improving software quality, with particular attention to large-scale, long-lived industrial systems and to programs written in an educational context. I investigate how programming faults, violations of coding conventions, and discrepancies between the intended and the actually implemented software architecture can be detected automatically from source code. I also work on model-based design of software systems and on automatic code generation from these models. An important application area of my research is the automated assessment of student assignments: relying on static analysis and software metrics, I measure the complexity, maintainability and correctness of submitted solutions and compare them with the characteristics of open-source projects. In addition, I am engaged in the methodological renewal of software engineering education, including the introduction of agile development methodologies and team-based, project-oriented teaching at university level.