Evaluating New Visualization Services for Browsing Spatiotemporal Earth Observation Satellite Data

Abstract:

Viewing raster imagery over web map tile services has been the standard method for rapidly browsing Earth observation satellite data. The Global Imagery Browse Services (GIBS), NASA’s premier image archive for viewing Earth observation data, has served billions of image tiles to science users and to the general public over the last several years with great success. However, there are current limitations to the types of data that can be served and visualized. For instance, three-dimensional data such as point clouds and data curtains simply do not work with standalone raster image tiles. Additionally, there are considerable costs associated with pre-processing and tiling imagery to provide users with a fast, lightweight experience. This research involves the evaluation and prototyping conducted by the GIBS team to develop new visualization services that can support these types of data with reasonable cost and performance at scale. We looked at methods for ingesting, storing, and serving these data as well as how they would be visualized on various clients including web applications, GIS software, and new platforms such as augmented and virtual reality. We also investigated the use of on demand image generation to reduce the need for pre-processing data into intermediary file formats while maintaining reasonable request latency. This includes visualization from higher precision cloud-optimized data—rather than PNG and JPEG images—which can enable additional image capabilities such as contrast stretching, dynamic color tables, and alternate band combinations.