Reduced Tolerance Imaging I
Abstract
The reduced tolerance imaging concept is to use imaging system hardware of reduced complexity to make phase-error-degraded measurements (or to lose phase information altogether) and then reconstruct diffraction-limited imagery in a post-detection processing stage using a phase retrieval algorithm. In the first year of a two-year effort several advances were made toward this end. An estimation theoretic (Cramer-Rao) lower bound on the error of estimating a coherent image for far-field (Fourier) intensity (squared modulus) measurements was derived for the case of Gaussian detector noise. Uniqueness of reconstruction from Fourier modulus assuming a priori known support was proven for a particular class of objects -- sampled objects whose support (the area outside of which it is zero) has a convey hull with no parallel sides. A closed- form recursive reconstruction algorithm was developed for reconstructing such objects via their autocorrelation functions. Simulations showed the closed-form solution to be sensitive to noise compared with iterative Fourier transform algorithms.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 01, 1986
- Accession Number
- ADA170932
Entities
People
- David L. Neuhoff
- Jack N. Cederquist
- James Fienup
- Richard G. Paxman
- Thomas R. Crimmins
Organizations
- Environmental Research Institute of Michigan