Superoverlay Deployment in Grid-Enabled Image Processing

SuperOverlays are image pyramid constructs for use in visualisation of client side imagery in the Google Earth (GE) virtual globe viewer. SuperOverlays combine portable image formats, such as JPEG, PNG or TIFF, with geographical metadata tagging in KML files. This has the fortunate side-effect that SuperOverlays can be easily served via standard web servers. Furthermore, it is relatively simple to programmatically access relevant parts of the SuperOverlay via their URI. This, then, allows the use of SuperOverlays in image processing, which can optionally be supported with a grid computing set-up. Presenting image processing output is facilitated by collecting grid node outputs in a SuperOverlay structure. In this paper, we demonstrate the use of a texture algorithm with SPOT-5 optical satellite data for benchmarking of the grid-enabled SuperOverlay processing against current practice single platform CPU processing. We discuss the impact of our solution on the feasibility for an image processing service configuration that can run in a client-server environment, possibly using GE as a user interface.