A Method for Specifying the Noise Suppression-Resolution Tradeoff in Digital Image Filtering with Local Statistics.

Abstract

A modification of Lee's local-statistics method is developed for digital image noise filtering without the need for a user-specified noise level. Instead, the user specifies an approximate upper size limit for image features he is willing to regard as noise. A noise level is then estimated automatically for each local region of the image by partitioning the region into subregions and exploiting the differing degrees of spatial correlation for signal and noise. Extensions for use with multiplicative noise, third-moment statistics, and edge-detection schemes are discussed briefly. Examples are given for an image of 256 x 256 pixels. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 11, 1980
Accession Number
ADA093560

Entities

People

  • Warren W. Willman

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Algorithms
  • Change Detection
  • Coordinate Systems
  • Detection
  • Digital Image Processing
  • Digital Images
  • Equations
  • Image Processing
  • Images
  • Probability
  • Probability Distributions
  • Radar
  • Radar Images
  • Random Variables
  • Standards
  • Statistics
  • Synthetic Aperture Radar

Fields of Study

  • Physics

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Computer Vision.
  • Statistical inference.