DocumentCode :
2208097
Title :
Moving Object Velocity Estimation Using Feature Point Image
Author :
Ejiri, Yusuke ; Hamada, Nozomu
Author_Institution :
Sch. of Integrated Design Eng., Keio Univ., Yokohama
fYear :
2006
fDate :
14-17 Nov. 2006
Firstpage :
1
Lastpage :
3
Abstract :
The analysis of moving object, such as extraction of moving object in three-dimensional space and its velocity estimation are widely studied in the fields of video surveillance system, robot vision, and efficient coding of image sequence. This paper is concerned with velocity estimation of moving objects, especially, spatiotemporal mixed-domain method (mixed method). Mixed method is known to be robust for noise and shape change of the moving object. But mixed method needs background subtraction because the estimation accuracy declines by the influence of background. Conventional background subtraction, such as background difference method or frame differencing method, cannot completely lost the influence of background. To solve the problem, we make feature point image and apply mixed domain method to it. The feature point image is image that only feature point has intensity. The proposed method realizes the high accuracy of velocity estimation
Keywords :
feature extraction; motion estimation; object detection; spatiotemporal phenomena; velocity measurement; background subtraction; feature point image; moving object analysis; spatiotemporal mixed-domain method; velocity estimation; Image analysis; Image coding; Image sequence analysis; Image sequences; Machine vision; Noise robustness; Orbital robotics; Robot vision systems; Spatiotemporal phenomena; Video surveillance;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
TENCON 2006. 2006 IEEE Region 10 Conference
Conference_Location :
Hong Kong
Print_ISBN :
1-4244-0548-3
Electronic_ISBN :
1-4244-0549-1
Type :
conf
DOI :
10.1109/TENCON.2006.343962
Filename :
4142552
Link To Document :
بازگشت