Stereo Camera Calibrations with Optical Flow

Abstract

RPA are currently unable to refuel mid-air due to the large communication delays between their operators and the aircraft. AAR seeks to address this problem by reducing the communication delay to a fast line-of-sight signal between the tanker and the RPA. Current proposals for AAR utilize stereo cameras to estimate where the receiving aircraft is relative to the tanker, but require accurate calibrations for accurate location estimates of the receiver. This paper improves the accuracy of this calibration by improving three components of it: increasing the quantity of intrinsic calibration data with CNN preprocessing, improving the quality of the intrinsic calibration data through a novel linear regression filter, and reducing the epipolar error of the stereo calibration with optical flow for feature matching and alignment. A combination of all three approaches resulted in significant epipolar error improvements over OpenCVsstereo calibration while also providing significant precision improvements.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2021
Accession Number
AD1132785

Entities

People

  • Joshua D Larson

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Air Platforms
  • Autonomy

DTIC Thesaurus Topics

  • Accuracy
  • Air Force
  • Aircrafts
  • Artificial Intelligence
  • Artificial Intelligence Software
  • Automata Theory
  • Computational Science
  • Computer Science
  • Computer Stereo Vision
  • Computer Vision
  • Computers
  • Convolutional Neural Networks
  • Deep Learning
  • Engineering
  • Image Segmentation
  • Information Systems
  • Line Of Sight
  • Machine Learning
  • Motion Capture
  • Neural Networks
  • Pattern Recognition
  • Refueling
  • Refueling In Flight
  • Stereo Cameras
  • Three Dimensional
  • United States
  • United States Government

Readers

  • Aerospace logistics and air mobility.
  • Geodesy
  • Neural Network Machine Learning.