Radial processing for feature detection of omnidirectional images

Jin Shin, Soo Yeong Yi

Research output: Contribution to journalArticlepeer-review

Abstract

Since omnidirectional images contain many circular arc segments, radial directional processing for detection of certain features in the image is more appropriate than horizontal or vertical directional processing, as with usual rectangular images. Computation of the proposed radial processing algorithm is based on the well-known Bresenham raster algorithm for line drawing in discrete integer space, and thus it is very fast and efficient. Experimental comparisons between the proposed algorithm and conventional horizontal processing in terms of feature detection and computational efficiency are presented.

Original languageEnglish
Pages (from-to)1567-1572
Number of pages6
JournalInformation (Japan)
Volume16
Issue number2 B
StatePublished - Feb 2013

Keywords

  • Bresenham raster algorithm
  • Image feature point
  • Omnidirectional structured light image

Fingerprint

Dive into the research topics of 'Radial processing for feature detection of omnidirectional images'. Together they form a unique fingerprint.

Cite this