【作者】 孙斌；

【导师】 张新荣；

【作者基本信息】 天津大学 ， 计算机应用技术， 2006， 硕士

【摘要】 本文研究的主要内容是图像序列中运动物体的检测与跟踪技术。图像序列中运动物体的检测与跟踪是计算机视觉领域的重要研究课题之一，也是近年来备受研究者关注的前沿方向,在智能监控系统、虚拟现实、高级用户接口、运动分析和基于模型的编码等领域都有应用。由运动物体所形成的图像序列可分为两种情况：一种是静止背景，一种是运动背景。本文主要研究静止背景下的运动物体的检测与跟踪算法。本文研究了目前运动物体检测的一些常用方法，包括基于帧间差分的方法、基于背景差分的方法。在基于背景差分的方法中，为了解决检测过程中背景光照亮度渐变和背景物体移动等问题，本文使用了一种基于动态信息窗口(DIW)的自适应背景更新算法。该算法根据像素的动态信息决定背景更新策略，并使用一种新的颜色模型有效地计算色度差，从而解决运动物体的阴影消除问题。利用该方法进行实验，实验结果表明所使用的方法能有效的检测出运动物体，而且能够消除阴影。本文也研究了运动物体跟踪的一些方法，包括基于扩展的卡尔曼滤波器的跟踪算法和基于颜色概率模型的CAMShift跟踪算法。通过采集的图像序列使用两种跟踪算法对人体跟踪进行实验，结果表明了跟踪算法的有效性。针对CAMShift跟踪算法存在的问题，给出了相应的解决方案。更多还原

【Abstract】 The thesis is focused on the technique of detection and tracking of the moving objects in image sequences. Detection and tracking of moving objects in image sequences is one of the most important research topics in the domain of computer vision, it is also an active field which has interested many researchers in recent years.It has applications in smart surveillance system,virtual reality,advanced user interface,motion analysis,model-based coding and so on. Image sequences of moving objects can be divided into two types by which background is static or dynamic.We major study detection and tracking algorithms of moving objects about the static background in this thesis.Some common approaches in the field of detection moving objects are studied,these approaches include the approach based on frame subtraction and the approach based on background subtraction in moment. For the approach based on background subtraction, to cope with these problems such as background illumination brightness incremental change and the moved background objects during the detection process, we use an adaptive background updating algorithm based on dynamic information window (DIW), the approach decided to update the strategy by the dynamic information of pixels.To cope with shadow problem of moving objects, a new computational color model is used. Chromaticity distortion is measured in an effective way. Using this method to do experiments, the experimental results show that the method can effectively detect moving objects, and to eliminate the shadow.Some approaches in the field of tracking moving objects are studied,these approaches include tracking algorithm based on extend Kalman filtering and CAMShift tracking algorithm based on colors probability model. By the collection of image sequences we use two kind of tracking algorithms do experiments about human body tracking, results show the effectiveness of tracking algorithms. Aimed at the problems of CAMShift tracking algorithms, the corresponding solution is given.更多还原