Enhancement of Target Contrast in Polarimetric Imagery Using Image Fusion

Abstract

We have examined the use of two statistical/numerical image fusion techniques for polarimetric imagery that each allow for an objective analysis of the resultant data products. The first technique is principal component analysis (PCA), a standard multivariate image processing method used to redistribute the information within the data set and possibly reduce the dimensions of the data set. The second method fuses conventional thermal imagery with a novel data product, the enhanced degree-of-linear polarization (DOLP). We performed receiver operating characteristic (ROC) curve analysis to quantify the resultant contrast between the manmade objects in the scene and the natural backgrounds. We found that the PCA data products inconsistently conveyed the desired contrast information, and that ROC curve analysis is not well suited to quantify the contrast in the PCA imagery. On the other hand, fusing the conventional thermal and the enhanced DOLP images provided a robust data product that effectively combines the information provided by the sensor.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 2010
Accession Number
ADA519582

Entities

People

  • Kristan Gurton
  • Melvin Felton

Organizations

  • United States Army Research Laboratory

Tags

Communities of Interest

  • Sensors

DTIC Thesaurus Topics

  • Cloud Cover
  • Clouds
  • Contrast
  • Data Science
  • Data Sets
  • Detectors
  • Factor Analysis
  • Image Processing
  • Information Processing
  • Information Science
  • Linear Polarization
  • Measurement
  • Polarization
  • Precision
  • Standards
  • Thermal Images

Readers

  • Computer Vision.
  • Image Processing and Computer Vision.
  • Regression Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference