Digital Library

cab1

 
Title:      IMPROVING THE ACCURACY OF POSITION DETECTION OF POINT LIGHT SOURCES ON DIGITAL IMAGES
Author(s):      Thilo Bauer
ISBN:      978-972-8924-84-3
Editors:      Yingcai Xiao, Tomaz Amon and Piet Kommers
Year:      2009
Edition:      Single
Keywords:      Image analysis, position detection, astronomy, astrometry.
Type:      Full Paper
First Page:      3
Last Page:      14
Language:      English
Cover:      cover          
Full Contents:      click to dowload Download
Paper Abstract:      The determination of accurate positions of stars is an essential task to obtain a reference coordinate system for precise navigation of ships, airplanes, satellites and space crafts. Stellar positions and proper motions will also help to investigate the structure and evolution of matter in the universe. From astronomical observations systematic errors have been reported by different authors using different techniques to detect and measure positions of point-light sources from digital images. An overview of this problem of image processing is given. The relative subpixel deviation will be defined and described. It provides interesting characteristics independent from the method of position retrieval or point spread introduced by any image deterioration. The relative subpixel deviation defines the lower limit of the achievable precision of positions. The function is the error correction term by itself, which yields a new concept of the error correction. A concept of a new simulation software is presented to further investigate the relative subpixel deviation. Features of the core framework are: support of different optical systems, sources of noise, and detectors types, like monochrome or color CCD and CMOS imagers, their pixel geometry, gaps and varying subpixel sensitivity functions. First results from the simulator software as well as results from new astronomical observations are presented as a proof of concept of the proposed approach to improve the current limit of position accuracy in the order of a few 1/100 of a pixel.
   

Social Media Links

Search

Login