Hendrik Herold explores potentials and hindrances of using retrospective geoinformation for monitoring, communicating, modeling, and eventually understanding the complex and gradually evolving processes of land cover and land use change. Based on a comprehensive review of literature, available data sets, and suggested algorithms, the author proposes approaches for the two major challenges: To address the diversity of geographical entity representations over space and time, image segmentation is considered a global non-linear optimization problem, which is solved by applying a metaheuristic algorithm. To address the uncertainty inherent to both the data source itself as well as its utilization for change detection, a probabilistic model is developed. Experimental results demonstrate the capabilities of the methodology, e.g., for geospatial data science and earth system modeling.

Hendrik Herold holds a doctoral degree from Dresden University of Technology, Germany, where he studied Geography, Geology, and Meteorology.

Monitoring and Modeling Land Change.- Geoinformation from Digital Images.- An Adaptive Map Image Analysis Approach.- Modeling Uncertainty for Change Analysis.