SVD Pseudo-Inverse Deconvolution of Two-Dimensional Arrays

Abstract

This thesis considers the expected matched filter response to a signal transmitted through a communications channel whose average scattering properties are known in terms of a scattering function. The matched filter is treated as an image which has been blurred by the properties of the interrogating signal. Removing this blurring is called deconvolution and is the problem addressed in this thesis. The problem is formulated to allow efficient application of the Singular Value Decomposition (SVD) as a method of deconvolution. It is shown that this form is the identical operation to the standard deconvolution via spectral division. Additionally, the problem of noise in the image is addressed and the trade-off between resolution and noise is discussed.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 26, 1985
Accession Number
ADA247767

Entities

People

  • M. A. Matuson

Organizations

  • Pennsylvania State University

Tags

Communities of Interest

  • C4I
  • Energy and Power Technologies
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Accuracy
  • Algorithms
  • Computational Complexity
  • Decomposition
  • Discrete Fourier Transforms
  • Dynamic Range
  • Eigenvalues
  • Eigenvectors
  • Equations
  • Filters
  • Frequency
  • Matched Filters
  • Navy
  • Random Variables
  • Scattering
  • Transfer Functions
  • Two Dimensional

Fields of Study

  • Engineering

Readers

  • Image Processing and Computer Vision.
  • Linear Algebra