Research on multi-pixels edge detecting and matching

Yu Liu, Yan Jun Li, Ke Zhang, You Yi Jiang

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

Abstract

Aiming at the influence of geometrical deformation on image matching, the method based on gray-level difference between the nearby pixels detecting multi-pixels edge of image is put forward. To a 256 by 256 image, it needs only 0.14 second to detect edge which is 40% and 35% faster respectively than that of the Prewitt and Robert arithmetic. Applying the method based on multi-pixels edge character to image matching, the influences of edge excursion can be improved obviously: when excursion exists in single pixel point, the relative probability of image matching is improved to 67%; when it exits in edge, it is improved to 50%. Research shows that this method is of simple model, high real-time quality and anti-geometrical-deformation. Most importantly, it is simply conducted.

Original languageEnglish
Title of host publication27th International Congress on High-Speed Photography and Photonics
DOIs
StatePublished - 2007
Event27th International Congress on High-Speed Photography and Photonics - Xi'an, China
Duration: 17 Sep 200622 Sep 2006

Publication series

NameProceedings of SPIE - The International Society for Optical Engineering
Volume6279 PART 2
ISSN (Print)0277-786X

Conference

Conference27th International Congress on High-Speed Photography and Photonics
Country/TerritoryChina
CityXi'an
Period17/09/0622/09/06

Keywords

  • Edge detecting
  • Image matching
  • Multi-pixels edge
  • Single pixel edge

Fingerprint

Dive into the research topics of 'Research on multi-pixels edge detecting and matching'. Together they form a unique fingerprint.

Cite this