Vision based method for the localization of intelligent vehicles in loose constraint area

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The localization is always an important research topic in the field of intelligent vehicle. This paper proposed a novel accurate localization method for intelligent vehicle navigation in loose constraint area (LCA) that uses only a single monocular camera. First, to eliminate the impact of the perspective effect and reduce the computational dimension, Harris corner feature points of the raw image are projected to the Inverse Perspective Image. Match them with feature point from the feature local map, using Normalized Cross-Correlation algorithm (NCC), calculate the optimal localization of vehicle using Random Sample Consensus algorithm (RANSAC) assisted Extended Kalman filter and then, update the feature local map. The proposed methodology is validated in the real world using an intelligent vehicle; it also has high position accuracy and robustness in the complex illumination.

Original languageEnglish
Title of host publicationProceedings - 2016 IEEE International Conference on Vehicular Electronics and Safety, ICVES 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages89-94
Number of pages6
ISBN (Electronic)9781509029334
DOIs
Publication statusPublished - 19 Aug 2016
Event2016 IEEE International Conference on Vehicular Electronics and Safety, ICVES 2016 - Beijing, China
Duration: 10 Jul 201612 Jul 2016

Publication series

NameProceedings - 2016 IEEE International Conference on Vehicular Electronics and Safety, ICVES 2016

Conference

Conference2016 IEEE International Conference on Vehicular Electronics and Safety, ICVES 2016
Country/TerritoryChina
CityBeijing
Period10/07/1612/07/16

Keywords

  • Extended Kalman filter
  • Intelligent vehicle
  • Loose Constraint Area
  • Navigation
  • Normalized Cross-Correlation algorithm (NCC)
  • Vision

Fingerprint

Dive into the research topics of 'Vision based method for the localization of intelligent vehicles in loose constraint area'. Together they form a unique fingerprint.

Cite this