Determining Virtual Practicality From Physical Stereo Vision Images and GPS

Abstract

Current research efforts for AAR at AFIT utilize Stereo Computer Vision to compute a relative pose between a tanker and receiver aircraft. Due to costs, time, and availability, it can be onerous to test these algorithms using actual AF aircraft. Our solution to this problem consists of using a 3D Graphics Engine to simulate AAR endeavors. However, the question then arises, Does the virtual world accurately represent the physical world? This can be explored by comparing a set of truth data to a similar set of virtual data. First, a set of truth data is collected using physical aircraft. Next, using the same flight path as that of the truth data, a set of virtual data is collected. Finally, a comparison of the physical and virtual data can provide information regarding how well the virtual world accurately represents the physical world, and if so, to within what margin of error? The results show that the virtual world roughly approximates the physical world but performs 2-6x better on average.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 19, 2020
Accession Number
AD1103216

Entities

People

  • Bradley S French

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Air Platforms
  • Space

DTIC Thesaurus Topics

  • Air Force
  • Aircrafts
  • Algorithms
  • Computer Science
  • Computer Stereo Vision
  • Computer Vision
  • Engineering
  • Flight Paths
  • Graphics
  • Refueling
  • Refueling In Flight
  • Simulations
  • Tanker Aircraft
  • Three Dimensional
  • United States
  • United States Government
  • Unmanned Aerial Vehicles

Fields of Study

  • Computer science

Readers

  • Aerospace logistics and air mobility.
  • Computational Modeling and Simulation
  • Educational Psychology

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • Space