Research Expertise and Interest
city and regional planning, landscape architecture and environmental planning, geographic information systems, database design and construction, spatial analysis, pattern recognition computational morphology
Research Description
Professor Radke teaches Geographic Information (GI) Science and related courses in both the Department of Landscape Architecture & Environmental Planning, and the Department of City & Regional Planning. He has led an effort to bring GI System technologies to the campus and promote their use in the Bay Area. Professor Radke is recognized for his contributions to pattern recognition, specifically the development of metrics and methods for characterizing spatial structure, association and relationships between objects embedded in the landscape. His metrics, embedded in graph theory, attempt to eliminate scale and density constraints common to many popular spatial metrics and produce a more sophisticated definition. The applied results of his research occur within Geographic Information Science and seek to solve real world problems common in planning and design. One such metric has brought us a powerful new tool for delineating boundaries and transition zones (or ecotones) in very complex heterogeneous distributions. This metric allows us to define boundaries and detect change in complex patterns in the landscape which up until now was not possible. Another of his metrics generalizes the notion of neighborly and provides a tool for defining an entire spectrum of neighborhoods where spatial relationships between objects (plants) grow more complex. Where measuring and delineating neighborliness was once difficult and often impossible, we can now easily grow, map and measure sophisticated neighborhoods. The success of Professor Radke's Spatial Decomposition Project will help designers and planners better measure, track and document spatial structure and change in complex landscapes where sophisticated sensors record and log spatial distributions of phenomena beyond human comprehension. As these sensors and computers evolve, our ability to measure minute and accurate changes in the landscape will rapidly increase producing a data rich environment. Metrics, like those developed by Professor Radke, look hopeful for characterizing the morphology of landscape. Professor Radke applies these methods to the field of Environmental Planning where he designs and constructs hazard models in attempts to predict and assess risk. One such study mapped the risk of firestorms in the East Bay Hills, another detected potential erosion threats to the coastal zone in St. John (USVI), and another automatically generated street centerlines in the City of Berkeley. Professor Radke speaks to and consults for local , regional and national governments on issues relating to GI Science, environmental monitoring, and data mining.
