Super Resolution Imagery From Multi-Frame Sequences with Random Motion

Abstract

A general method is described for achieving super resolution imagery from multiple frame image sequences that contain motion. The method assumes low-noise, focal plane array imagery recorded with uncontrolled image motion that can included some random jitter. Using this approach, moderate resolution, fast frame image sequences can be processed to achieve high-resolution image sequences displayed at conventional frame rates. The super resolution processing depends only on the imagery, requiring no externally controlled micro-dither or a priori information such as the sensor motion or range to the background. Typical sensor stabilization requirements are relaxed using this method, however, to achieve optimum performance there are some constraints on the motion. Specifically, the stabilization must still good enough so that the resulting random dither is only a few pixels and must be statistically well behaved. Processing examples are given using previously recorded image sequences from a wide FOV MWIR staring array sensor on-board an aircraft.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 02, 1998
Accession Number
ADA400605

Entities

People

  • D. Scribner
  • J. Schuler
  • M. Kruer

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Air Platforms
  • Materials and Manufacturing Processes
  • Sensors

DTIC Thesaurus Topics

  • Aircrafts
  • Arrays
  • Detectors
  • Digital Image Processing
  • Digital Images
  • Distortion
  • Filters
  • Focal Plane Arrays
  • Focal Planes
  • Frequency
  • Frequency Domain
  • Image Processing
  • Image Restoration
  • Images
  • Noise
  • Optical Images
  • Sequences

Readers

  • Control Systems Engineering.
  • Image Processing and Computer Vision.