Representations and topology-based methods for geospatial data analysis and visualization

Enabling insight into large and complex spatial datasets is the central theme in scientific data visualization research. Visualizing large spatial data sets requires efficient data representation methods, powerful analysis algorithms for data transformation and effective rendering and human-computer interaction methods to allow domain experts to gain insight from the data. We consider here geospatial data representing continuous phenomena, such as scalar fields (terrains, 2D or 3D images, unstructured volume data sets, etc.), and multifields, collections of fields with different modalities. The talk will focus on the first two steps of the data visualization pipeline, and specifically on compact and scalable representations for large-size spatial data sets, and on data transformation methods based on topological data analysis. Applications will be discussed to bathymetric data analysis for marine navigation, to tree segmentation from LiDAR (Light Detection and Ranging) point clouds for forest management, to multifield environmental data visualization.