III:Small:RUI:Integrating Image and Non-Image Geospatial Data

University Of California - Merced, Merced CA

Investigators

Abstract

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5). This project develops novel methods for integrating image and non-image geospatial data to (1) advance the state-of-the-art of automated remote sensed image analysis and, in turn, (2) improve the coverage and fidelity of the non-image repositories. A characteristic of remote sensed imagery which has not been sufficiently exploited by the analysis community is that the images can be georeferenced to extensive repositories of non-image geospatial data such as maps and geographic dictionaries termed gazetteers. In particular, these associations represent a rich source of labeled data needed to train the analysis algorithms. The first part of this project develops a framework for learning appearance models for a large set of geospatial objects indexed by an extensive gazetteer in an unsupervised fashion. Besides the standard challenges such as choice of features and form of the model, this problem is made interesting by the fact that current gazetteers only specify the spatial footprint of indexed objects using a single point location. Methods are explored for simultaneously estimating the model parameters and the spatial extents of the known objects. The second part investigates using the learned models to update the gazetteer from imagery. This includes estimating the spatial extents of known object instances as well as detecting unknown or novel instances. The project has research and educational synergies with a Spatial Analysis Research Center that the PI is establishing at the new University of California at Merced. The PI plans to open imagery so that evaluation datasets can be made publicly available through the project website (http://vision.ucmerced.edu/projects/integrating/).