92. PLATFORM MOTION BLUR IMAGE RESTORATION SYSTEM

Department: Electrical & Computer Engineering
Research Institute Affiliation: Graduate Program in Computational Science, Mathematics, and Engineering (CSME)
Faculty Advisor(s): Joseph Ford

Primary Student
Name: Stephen Joseph Olivas
Email: sjolivas@ucsd.edu
Phone: 858-822-4406
Grad Year: 2013

Abstract
Restoring images degraded by platform motion blur is a common problem for space based astronomical imagers. Platform motion blur, blur caused by a moving camera, is accentuated by long exposure times and large object distances. It also has structure in that the entire image moves as a rigid object. One general restoration approach is to deconvolve the image with a point spread function (PSF) which describes the blur. The blur is in general spatially variant (SV), meaning it is defined differently for different parts of the image. Since platform motion is structured, it is possible to calculate the SV PSF using at least three local PSF measurements made at different locations within the image. This prompted the fabrication of a custom prototype system capable of acquiring images while simultaneously tracking image motion at specific image locations. These motion measurements are made using position sensing detectors (PSD). The PSDs are photodiodes capable of accurately tracking bright image features that appear on a dark background, making them well suited for starfield imaging. We form a SV PSF from the PSD data our prototype provides in order to significantly restore blurred images using a pixel-by-pixel deconvolution method.

« Back to Posters or Search Results