运动目标光流场算法研究进展  

Development of algorithm of optical flow field

作　　者：黄志良[1] 张利勋[1] 

机构地区：[1]电子科技大学光电信息学院,成都610054

出　　处：《激光杂志》2013年第1期1-3,共3页Laser Journal

摘　　要：光流场技术是一种全新的用于目标跟踪的技术,主要用于机器视觉识别,自动飞行器导航和天气预报监测等。与传统的技术相比,光流场技术的优势主要是目标分割的简易性,实时性和准确性。图像中的运动分析在计算机视觉中占有十分重要的地位,所以光流场已成为热点研究领域之一。本文首先介绍了光流场算法的演变历史,然后分析当今世界几种先进的算法,如内插值法的光流场4D CT图重构,光流场主动轮廓法的图像序列分割,光流场全方位摄影法的成像分割,以及高速物体运动下对成像误差的影响,指出了当前我国研究工作所面临的问题和今后的发展趋势。The technology of Optical Flow Field is a new technology as target tracking.It is used in machine visual identity,automatic aircraft navigation,weather forecast monitor and so on.Comparing with traditional technology,the advantage of technology of Optical Flow Field is most the simplicity,real-time and accuracy of target segmentation.Through reading the literature of Algorithm of Optical Flow Field of home and abroad,sums up the history of Algorithm of Optical Flow Field,analyses the principle,characteristic and result of some advanced algorithm of current society.For example,interpolation method of optical flow field 4D CT image reconstruction,active contours method of image sequence segmentation,omnidirectional camera image method of image sequence segmentation.And the influence of image error in the situation of the moving aim with high speed.Point out the problem which we should confront and development tendency in future.

关 键 词：光流场 4D CT图 主动轮廓 全方位摄影 成像误差 

分 类 号：TN248.1[电子电信—物理电子学]