Extreme Scale Astronomical Image Composition and Analysis

Robert J. Brunner, University of Illinois at Urbana-Champaign

Usage Details

Eric Shaffer, Robert J. Brunner, Britton Jeter, Harshil Kamdar, Ching Chang, Sean McLaughlin, Matias Carrasco Kind, Matias Carrasco Kind

We propose to take the small, calibrated images from the Sloan Digital Sky Survey (SDSS) (and Dark Energy Survey as they become available), and stitch them together into calibrated science-grade images. By doing this, we will be uniquely positioned to mine these rich image data for large-scale features such as unknown nearby dwarf galaxies or tidal streams of stars resulting from interactions between our own Milky Way galaxy and local group galaxies. We have used the open-source package montage to create image mosaics that are roughly one-square degree across, which requires over 100 individual SDSS field images to be registered, aligned and stitched into a new resulting image, that were made through this process; the color image is made by combining images in three filters together in a second stage process).  Our ultimate goal is to make a single stitched image of the entire SDSS survey in each band.

Each of these images would exceed one terapixels, making their creation and subsequent analysis an extremely large computational challenge, ideally suited to Blue Waters large disks system, large memory nodes, and on-board GPU processors.

http://lcdm.astro.illinois.edu/