A Particle Method for Finding Distributed Objects

Abstract

Detecting and identifying distributed objects in an image is a challenge with wide research application. This search problem is difficult, in part, because the individual, multipixel "spots" that make up a distributed object must be taken in the aggregate to have any relevance, and identification is possible only when the majority of the individual spots are detected and found to conform to an expected pattern. In this paper, particle filtering methods are extended in order to detect, localize, and identify a distributed object in a single cluttered image by maximizing the joint probability that a particular collection of spots is the object of interest. The method is illustrated using a "surrogate" estimation problem. Results demonstrate that the proposed method gives a high probability of correct detection and low object location error when the signal to clutter-plus-noise ratio is above 5 decibels.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 2004
Accession Number
ADA524344

Entities

People

  • John H. Greenewald
  • Stanton H. Musick
  • Teri L. Piatt

Organizations

  • Air Force Research Laboratory

Tags

Communities of Interest

  • Air Platforms
  • Biomedical
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Air Force
  • Air Force Research Laboratories
  • Algorithms
  • Computer Vision
  • Detection
  • Detectors
  • False Alarms
  • Filtration
  • Geometry
  • Heuristic Methods
  • Identification
  • Navigation
  • Object Recognition
  • Particles
  • Sampling
  • Sequential Monte Carlo Methods
  • Warning Systems

Fields of Study

  • Computer science

Readers

  • Computer Vision.
  • Radar Systems Engineering.
  • Systems Analysis and Design