The Effect of Variability in Stereo Camera Calibration and Registration Parameters on Three-Dimensional Reconstruction Distance Estimates

Abstract

The process of calculating distance front binocular or stereo vision is tented three-dimensional (3-D) reconstruction. Unlike some other approaches for estimating distance, e.g., LADAR (laser detection and ranging), this is not a direct measurement but is a geometric analysis requiring the knowledge of experimentally measured and mathematically derived parameters. Variability in these parameters results in variability in the calculated distances. This report investigates the effect that the variability in these parameters has on computed 3-D reconstruction distances. Results indicate that sub-pixel and milliradian accuracy are required for the parameters in order to have less than a 10% variation in the 3-D reconstruction process.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Feb 01, 2004
Accession Number
ADA421239

Entities

People

  • William F. Oberle

Organizations

  • United States Army Research Laboratory

Tags

Communities of Interest

  • Air Platforms
  • Autonomy
  • Sensors
  • Weapons Technologies

DTIC Thesaurus Topics

  • Accuracy
  • Calibration
  • Cameras
  • Computer Stereo Vision
  • Computer Vision
  • Coordinate Systems
  • Detection
  • Detectors
  • Image Processing
  • Jet Propulsion
  • Laser Detection
  • Laser Radar
  • Laser-Based Detection
  • Measurement
  • Stereo Cameras
  • Three Dimensional
  • Unmanned Ground Vehicles

Fields of Study

  • Physics

Readers

  • Computer Vision.
  • Regression Analysis.

Technology Areas

  • Directed Energy