Authors: Timothy Schempp*, San Diego State University, André Skupin, San Diego State University
Topics: Geographic Information Science and Systems, Spatial Analysis & Modeling, Land Use and Land Cover Change
Keywords: image analysis, multivariate, spatiotemporal analysis, remote sensing, land cover change, dimensionality reduction, spatialization
Session Type: Paper
Start / End Time: 8:00 AM / 9:40 AM
Room: Southdown, Sheraton, 4th Floor
Presentation File: No File Uploaded
Continued development of computational methods and user interfaces that explicitly visualize change associated with spatiotemporal phenomena supports applications in many domains. These include epidemiology, law enforcement, precision agriculture, environmental monitoring, urban studies, and many more. The tri-space framework provides a means of analyzing multivariate and multitemporal data by manipulating its structure to yield six distinct perspectives, each illiciting unique insights on the data. Dimensionality reduction techniques, such as self-organizing maps and multidimensional scaling, facilitate interpretation through visualization. Imagery analysis presents a compelling domain for this methodology as satellite imagery is commonly used in a wide variety of environmental and urban studies in order to monitor dynamic phenomena. Unlike conventional approaches to understanding land cover change, the tri-space framework enables the visualization of pixel trajectories across multispectral space. Additionally, holistic trends and more subtle patterns present within the data can both be detected and emphasized by the tri-space framework and the different normalizations it provides. Understanding can be further enhanced through classification overlays, interactivity, and the linked visualization between tri-space perspectives and geographic space. The research presented here focuses on the conceptualization and scalable development of a web-based platform that encapsulates the tri-space approach in an exploratory data analysis environment to facilitate inductive investigation of multitemporal imagery data.