We present an efficient method based on image rectification for achieving the feature correspondence between images. The method is comprehensible through a geometric analysis of two rectified views. An important constraint-Truncation Point-is found and largely reduces a search range than the conventional method. The classical normalized cross-correlation is then performed on this search range to locate the correspondence points and tracking is followed further. The experiment demonstrates the considerable improvements obtained by this method as well as the good feature tracking effect.
Published in:
Information and Computing (ICIC), 2010 Third International Conference on
(Volume:3
)
Date of Conference: 4-6 June 2010