New Algorithms for the Reduction of Image Degradation Due to Atmospheric Turbulence

Abstract

The problem of removing the deleterious effects of clear-air turbulence from images has been formally solved. As a spinoff, the method also allows wind velocity/shear to be measured along the optical line of sight. The overall approach uses two short-exposure images as inputs. These are divided in Fourier space. The division data is independent of the object (since it is the same in both images), and depends upon parameters that specify the two point spread functions. A neural net allows these parameters to be found, with good accuracy. The two images are then inverse-filtered, using the p.s.f. information. The average of the two inverse-filtered outputs is then the final reconstruction. Because of the neural net implementation, the entire algorithm can be emplaced upon a logic chip and used in quasi-real time. Future development would seem to promise real-time operation. Image reconstruction, Clear air turbulence, Smart camera

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 22, 1994
Accession Number
ADA284891

Entities

People

  • B. R. Frieden

Organizations

  • University of Arizona

Tags

Communities of Interest

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

DTIC Thesaurus Topics

  • Algorithms
  • Atmospheric Motion
  • Clear Air Turbulence
  • Digital Image Processing
  • Digital Images
  • Image Processing
  • Image Reconstruction
  • Neural Networks
  • Power Spectra
  • Scientists
  • Spectra
  • Turbulence
  • Wind
  • Wind Shear
  • Wind Velocity

Fields of Study

  • Physics

Readers

  • Image Processing and Computer Vision.
  • Mathematical Modeling and Probability Theory.

Technology Areas

  • Space