活动图象的锥体分析法

Image Motion Analysis with the Pyramid

摘要: 本文提出了一种新的活动图象分析法，该方法建立在高效的多级分辨率图象锥体算法的基础上，通过相关匹配估计图象序列连续帧间的位移，给出了计算机模拟结果．

Abstract: A new method for image motion analysis is proposed.It is based on the multi-resolution pyramid algorithms with high efficiency and uses correlation mathing process to estimate the displacement between the motion sequence frames.The analysis is decomposed into four distinct stages and independent channels,so it is possible to perform the computation in parallel pipelined hard ware,Some computer simulation results are given.

[1]	Barnea,D.and Silverman,H.,A,Class of Algorithms for Fast Digital Image Registrstion,IEEE Trans.Computer,179-186,1972.
[2]	Frei,W.et al.,Fast Matching of Non-Stationary Images with False Fix Protection,Proc.5th Intl.Conf.on Pattern Recognition,208-212,1980.
[3]	Wong,R.and Ha11,E.,Sequential Hierarchical Scene Matching,IEEE Trans.Computers,359-366,1978.
[4]	Burt,P.,Fast Filter Transforms for Image.Processing,Compater Graphies and Image Processing,16,20-51,1981.
[5]	Adelson,E.and Burt,P.,Image Data Compressioa with the Laplacian Pyramid,Proc.IEEE Conf.on Pattern Recognition sad Image Processing,218-223,1981.
[6]	Burt,P.,Yen.C.and Xu.X.,Local Correlation Measures for Motion Analysis.Proc.IEEE Conf.on Pattern Recognition and Image Processing,269-274,1982.
[7]	Burt,P.,Yen,C.and Xu.,X.Multi-resolution Flow-through Motion Analysis,Proc.IEEE Conf.on Computer Vision and Patters Recognition.246-252,1998.