A segmentation method for multi-connected particle delineation

Xingqiang Wu, John M. Kemeny

Research output: Chapter in Book/Report/Conference proceedingConference contribution

22 Scopus citations

Abstract

An automatic particle segmentation system is developed for calculating the size distribution of rock fragments created by blasting. A rock composite due to blasting is often fully multi-connected in which individual particles cannot be delineated by the existing segmentation algorithms. Two algorithms are proposed to approach this multi-connected segmentation problem. The first algorithm analyzes the shape of each shadow (a simply-connected region) and "splits" the particles from shadow boundary convexity points if a relatively large gradient path occurs. The second algorithm finds clusters of rock particles which may not be delineated due to the lack of a strong gradient along the touching portions and delineates them using a shape heuristics. A large number of test results show that the method is fast and accurate.

Original languageEnglish (US)
Title of host publicationIEEE Workshop on Applications of Computer Vision
PublisherIEEE Computer Society
Pages240-247
Number of pages8
ISBN (Electronic)0818628405
DOIs
StatePublished - 1992
Event1992 IEEE Workshop on Applications of Computer Vision, WACV 1992 - Palm Springs, United States
Duration: Nov 30 1992Dec 2 1992

Publication series

NameProceedings of IEEE Workshop on Applications of Computer Vision
Volume1992-November
ISSN (Print)2158-3978
ISSN (Electronic)2158-3986

Conference

Conference1992 IEEE Workshop on Applications of Computer Vision, WACV 1992
Country/TerritoryUnited States
CityPalm Springs
Period11/30/9212/2/92

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Computer Science Applications

Fingerprint

Dive into the research topics of 'A segmentation method for multi-connected particle delineation'. Together they form a unique fingerprint.

Cite this