Skip to main navigation Skip to search Skip to main content

Train wheel edge detection and image object region segmentation

  • Guo Nan
  • , Shengfang Lu
  • , Junen Yao
  • Beihang University

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

Abstract

Machine vision based mechanical appearance fault analysis and inspection is getting broad applications in past decades. Train wheel tread damage is a common fault pattern. The precedent step of the routine vision based analysis work is to get an image that includes the wheel surface. In this paper, a wheel curve edge extraction and object region segmentation framework is proposed. Firstly the salient rail line edge is extracted for a previous segmentation step and a sub image is acquired. Then line segment detector is used to detect the lines along the contours. And the wheel and shadow curve edge are approximated by line segments sets. Through certain geometry rules, the two edge lines are extracted. Finally the wheel object region is extracted perfectly and accurately.

Original languageEnglish
Title of host publicationInfrared Technology and Applications, and Robot Sensing and Advanced Control
EditorsHaimei Gong, Aiguo Song
PublisherSPIE
ISBN (Electronic)9781510607729
DOIs
StatePublished - 2016
EventInternational Symposium on Infrared Technology and Application and the International Symposiums on Robot Sensing and Advanced Control - Beijing, China
Duration: 9 May 201611 May 2016

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume10157
ISSN (Print)0277-786X
ISSN (Electronic)1996-756X

Conference

ConferenceInternational Symposium on Infrared Technology and Application and the International Symposiums on Robot Sensing and Advanced Control
Country/TerritoryChina
CityBeijing
Period9/05/1611/05/16

Keywords

  • Object region segmentation
  • Rail line extraction
  • Train wheel edge detection
  • Visual inspection system

Fingerprint

Dive into the research topics of 'Train wheel edge detection and image object region segmentation'. Together they form a unique fingerprint.

Cite this