Image Processing Application for Cognition (IPAC) - Traditional and Emerging Topics in Image Processing in Astronomy

摘要

A new application framework for advanced image processing forudastronomy is presented. It implements standard two-dimensional operators,udand recent developments in the field of non-astronomical image processing (IP),udas well as original algorithms based on nonlinear partial differential equationsud(PDE). These algorithms are especially well suited for multi-scale astronomicaludimages since they increase signal to noise ratio without smearing localized anduddiffuse objects. The visualization component is based on the extensive tools thatudwe developed for Spitzer Space Telescope’s observation planning tool Spot andudarchive retrieval tool Leopard. It contains many common features, combines images in new and unique ways and interfaces with many astronomy data archives.udBoth interactive and batch mode processing are incorporated. In the interactiveudmode, the user can set up simple processing pipelines, and monitor and visualize the resulting images from each step of the processing stream. The systemudis platform-independent and has an open architecture that allows extensibilityudby addition of plug-ins. This presentation addresses astronomical applicationsudof traditional topics of IP (image enhancement, image segmentation) as well asudemerging new topics like automated image quality assessment (QA) and feature extraction, which have potential for shaping future developments in theudfield. Our application framework embodies a novel synergistic approach basedudon integration of image processing, image visualization and image QA (iQA).